Science + Tech – The Conversation

How AI is challenging the credibility of some online courses

2025-11-04T16:16:57Z

Distance learning far precedes the digital age. Before online courses, people relied on print materials (and later radio and other technologies) to support formal education when the teacher and learner were physically separated.

Today, there are varied ways of supporting distance learning with digital communication. With “asynchronous” online courses, teaching does not occur live. Students access course materials on the learning management system and complete assignments at their own pace. This allows flexibility across time zones and work schedules and affords accessible learning.

Nevertheless, some researchers have raised concerns regarding the quality and student outcomes associated with asynchronous online courses. As well, generative artificial intelligence (GenAI) has exposed fundamental challenges to this mode of delivery.

While GenAI poses serious challenges to academic integrity in many formats of learning, including synchronous online and in-person learning, asynchronous courses face the most acute risk. Without real-time interaction or time constraints, students can use AI undetected while instructors never observe their thinking processes.

Compromised learning models

Asynchronous courses have long relied on conventional assessments: discussion board posts, written reflections, essay assignments and pre-recorded videos. These models in asynchronous assessment are now compromised. Distinguishing AI-generated content from human-written text has become increasingly difficult.

Text discussions and reflections present the highest substitution risk. GenAI can generate personalized reflective posts and discussion replies rapidly, complete with a polished academic tone. An instructor may spend hours responding to these contributions yet gain little evidence about who actually learned and produced the material.

AI agents like ChatGPT’s Atlas browser can now navigate course sites, consume materials and complete some assignments with minimal student intervention — if any.

In written assignments, requiring precise citations from assigned course materials may seem like a safeguard. However, AI-enhanced tools can easily meet such requirements. This approach provides false security and fails to address underlying problems, including authorship with integrity in an AI world.

Students can be asked to provide time-stamped drafts, version history and checkpoints to document their process. But these can be easily fabricated — while instructors become overloaded with policing rather than focusing on students’ learning and progress.

AI-generated infographics and videos are also becoming hard to distinguish from human-made ones.

Detectors, remote proctoring not solutions

AI detectors cannot solve the problem. Research suggests detection tools produce false positive rates far higher than advertised, with disproportionate harm to neurodivergent and second-language learners. Several universities now explicitly advise against using detection software as evidence of academic misconduct.

Remote proctoring is intrusive and raises serious ethical, equity, privacy and reliability concerns. Students requiring accommodations, whether for disabilities, inadequate technology or lack of private space, must be granted leniency that undermines the system’s purpose, rendering it unsustainable while diverting instructors away from their educational mission.

These mounting challenges are neither hypothetical nor distant. Without meaningful intervention, institutions risk credentialing students who have not demonstrably engaged with course content, thereby undermining the integrity of academic credentials.

Two less-than-ideal strategies

Genuine protection against AI substitution requires approaches that fundamentally alter how instructors deliver asynchronous courses. Two strategies that somewhat meet this threshold are:

1) Short oral examinations can be scheduled for major assignments or throughout a term. While not without limitations, these conversations verify authorship and assess the depth of understanding.

2) Experiential learning components with external verification: Students can apply course concepts to real-world settings and include brief attestations from workplace supervisors, community partners or other external stakeholders in their capstone assignments that will be graded by course instructors. Combined with short oral examinations, this approach would deter offloading all learning to AI and augment asynchronous coursework with practical components.

However, assessment strategies alone cannot solve the authenticity crisis.

Rethinking program design

The Community of Inquiry framework, a tool for conceptualizing online learning, identifies three essential elements for effective online learning: social presence (students engage authentically), cognitive presence (students construct understanding through inquiry) and teaching presence (instructors facilitate learning).

GenAI threatens all three of these elements: it can simulate social engagement through generated posts, substitute for cognitive work and force instructors to focus on policing rather than teaching.

Institutions must evaluate whether their asynchronous programs can maintain these elements given GenAI capabilities.

Confronting an uncomfortable reality

Institutions and educators must be honest about limitations. Few strategies provide genuine protection against AI substitution; most merely create friction that determined students can overcome. The recommended approaches named above require synchronous elements or external verification that fundamentally alter asynchronous delivery.

Implementation of these imperfect solutions requires genuine institutional commitment, resources and policy support. Institutions now face a choice: invest substantially in what is required to restore some degree of assessment authenticity or acknowledge that asynchronous programs (as currently structured) cannot credibly assure learning outcomes.

Band-aid solutions and deflection of responsibility to instructors will only deepen the credentialing crisis. In the absence of robust institutional efforts, asynchronous programs risk becoming credential mills in all but name. The question is not whether institutions can afford to act, but whether they can afford not to.

Rahul Kumar received funding from SSHRC in the past.

Mohammed Estaiteyeh does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

How the physics of baseball explains Blue Jay Kevin Gausman’s signature pitch

2025-10-30T22:12:06Z

There are few sports more exciting than playoff baseball, but behind every pitch there is also a fascinating story of physics. From gravity to spin, the science shaping the game can be just as compelling as the action on the field.

When the World Series returned to Toronto for Game 6, right-handed pitcher Kevin Gausman took the mound for the first six innings. Gausman’s best pitch is the splitter, an off-speed pitch that looks like a conventional fastball but travels more slowly and drops more sharply before it crosses the plate.

Physicists consider the flight of a baseball as an example of projectile motion. The trajectory of the ball depends on several forces: the force of gravity (pulling the ball downwards), the drag force (slowing the ball as it moves through the air), and the Magnus force (which causes the ball to curve if it spins as it travels).

Why splitters are so hard to hit

So why is the splitter so difficult to hit? Start with speed. The average speed of Gausman’s fastball is 95 miles per hour (or 42.5 meters a second). Since the distance from the pitcher’s mound to home plate is 18.4 meters, this means that it takes 430 milliseconds, or less than half a second, for Gausman’s fastball to reach the batter.

In contrast, the splitter, which travels at an average speed of 85 mph (or 38.0 m/s), takes 490 milliseconds. That 60 millisecond-difference may seem small, but it can be enough to separate a strike from a base hit.

For context, a typical swing for a major league batter takes approximately 150 milliseconds. This includes time for the batter’s eye to form a picture of the ball leaving the pitcher’s hand, for their brain to process this information and send signals to the muscles in their arms, legs and torso, and for their muscles to respond and swing the bat.

This means a batter has roughly a quarter of a second to judge the trajectory of a pitch and decide whether to swing. Considering that it takes approximately 100 milliseconds for a blink of the human eye, it’s remarkable that batters can hit any major league pitch at all.

The importance of the drop

The second secret to the splitter is the drop. All baseball pitches drop as they travel towards home plate due to the force of gravity, which causes a baseball (or any object in freefall) to accelerate downwards.

If there were no other forces acting on the ball, this would cause Gausman’s fastball to drop by about 92 centimetres on the way to home plate, and his splitter to drop by approximately 115 centimetres.

In practice, however, there is another important force that acts on the ball to oppose the effect of gravity — the Magnus force. The Magnus force arises from the rotation or spin of an object (like a baseball) as it passes through a fluid (like air).

The ball’s rotation makes air move faster over one side than the other. On the side spinning in the same direction as the airflow, air speed increases; on the opposite side, it slows down. This difference in air speed creates a pressure imbalance, generating a force that acts perpendicular to the ball’s path.

This is an example of Bernoulli’s Principle, the same phenomenon that generates lift as air passes around the wing of an airplane.

In the case of a fastball, the pitcher creates a strong backspin by pulling back with their index and middle fingers as they release the ball. This rotation results in an upwards force, which causes the ball to drop far less than it would under the effect of gravity alone. The faster the rotation, the stronger this lift force becomes.

Gausman’s signature pitch

Gausman’s fastball typically drops 25 to 30 centimetres on the way to home plate — less than one third of the drop experienced by a “dead ball” without spin.

On the splitter, he changes his grip to dramatically reduce the amount of backspin, weakening the Magnus force and allowing the ball to fall much farther, about 50 to 75 centimetres, before it hits the plate. The result is a pitch that doesn’t reach the batter when or where they expect it to be.

Kevin Gausman explains the art of the splitter. (Toronto Blue Jays)

While the Blue Jays failed to win a third World Series title, Gausman’s splitter offered an example of how physics can shape performance in elite sport. Understanding the science behind the pitch offers a new way to appreciate the game.

Patrick Clancy does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Rate my AI teacher? Students’ perceptions of chatbots will influence how they learn with AI

2025-10-30T18:13:08Z

A “transformation” is upon us. After a multi-year procession of educational technology products that once promised to shake things up, now it’s AI’s turn.

Global organizations like the Organization for Economic Co-operation and Development, as well as government bodies, present AI to the public as “transformative.”

Prominent AI companies with large language model (LLM) chatbots have “education-focused” products, like ChatGPT Education, Claude for Education and Gemini in Google for Education.

AI products facilitate exciting new ways to search, present and engage with knowledge and have sparked widespread interest and enthusiasm in the technology for young learners. However, there are crucial areas of concern regarding AI use such as data privacy, transparency and accuracy.

Current conversations on AI in education focus on notions it will upend teaching and learning systems in schools, teacher lesson planning and grading or individualized learning (for example, via personalized student tutoring with chatbots). However, when or whether AI will transform education remains an open question.

In the meantime, it is vital to think about how student engagement with chatbots should make us examine some fundamental assumptions about human learning.

Learning is a social affair

How students view their teachers and their own ability to contemplate thinking (known as metacognition) are tremendously important for learning. These factors need to be considered when we think about learning with chatbots.

The popularity of the Rate My Professors website in Canada, United States and the United Kingdom is a testament to the significance of what students think about teachers.

With AI’s foray into education, students’ conceptions of their AI tutors, teachers and graders will also matter for multiple reasons.

First, learning is a thoroughly social affair. From how a child learns through imitating and modelling others to engaging with or being influenced by peers in the classroom, social interactions matter to how we learn.

With use of chatbots increasing to more than 300 million monthly users, conversational interactions with LLMs also represent a new para-social interaction space for people worldwide.

What we think of interaction partners

Second, theory-of-mind frameworks suggest that what we think of others influences how we interact with them. How children interpret, process or respond to social signals influences their learning.

To develop this idea further, beyond other students or teachers as interaction partners, what we think about learning tools has an influence on how we learn.

Our sense of tools and their affordances — the quality or property of a tool that “defines its possible uses or makes clear how it can or should be used” — can have consequences for how we use the tool.

Perceived affordances can dictate how we use tools, from utensils to computers. If a learner perceives a chatbot to be adept at generating ideas, then it could influence how they use it (for example, for brainstorming versus editing).

New ‘social entity’

AI systems, at a minimum, represent the entrance of a new social entity in educational environments, as they have in the social environment. People’s conceptions of AI can be understood under the larger umbrella of a theory of artificial minds, referring to how humans infer the internal states of AI to predict actions and understand behaviour. This theory extends the notion of theory of mind to non-human AI systems.

A person’s theory of artificial minds could develop based on biological maturation and exposure to the technology, and could vary considerably between different individuals.

3 aspects to consider

It’s important to consider how student conceptions of AI may impact trust of information received from AI systems; personalized learning from AI; and the role that AI may have in a child’s social life:

1. Trust: In human learning, the judgments we make about knowledge and learning go a long way in acceptance of ideas inherent in learning material.

From recent studies in children’s interactions with conversational AI systems, we see that children’s trust in information from AI varies across factors like age and type of information. A learner’s theory of artificial minds would likely affect willingness to trust the information received from AI.

2. Personalized learning: Intelligent tutoring systems (ITS) research has shown excellent results for how traditional ITS — without chatbot engagement — can scaffold learners while also helping students identify gaps in learning for self-correction. New chatbot-based ITS, such as KhanMigo from Khan Academy, are being marketed as providing personalized guidance and new ways to engage with content.

A learner’s theory of artificial minds could affect the quality of interactions between them and their AI chatbot tutor and how much they accept their learning support.

3. Social relationships: The artificial friend (the “AF”) in Kazuo Ishiguro’s Klara and the Sun is a poignant literary example of the impact an artificial entity can have on a growing child’s sense of self and relationship to the world.

We can already see the detrimental effects of introducing children to AI social chatbots with the tragic suicide of a child who was allegedly engaged in emotional and sexual chat conversations with a Character.AI chatbot.

Social relationships with AI involve a serious renegotiation of the social contract regarding our expectations and understanding of each other. Here, relationships with children need special attention, foremost whether we want children to develop social relationships with AI in the first place.

Where do we go from here?

Many discussions about AI literacy are now unfolding, involving, for example, understanding how AI functions, its limitations and ethical issues. Throughout these conversations, it’s essential for educators to recognize that students possess an intuitive sense of how AI functions (or a theory of artificial minds). Students’ intuitive sense of AI shapes how they perceive its educational affordances, even without formal learning.

Instruction must account for students’ cognitive development, existing experiences and evolving social contexts.

The “rate my AI teacher” future is coming. It will require a focus on students’ conceptions of AI to ensure effective, ethical and meaningful integration of AI into future educational environments.

Nandini Asavari Bharadwaj receives funding from Social Sciences and Humanities Research Council of Canada.

Adam Kenneth Dubé receives research funding from Mitacs, Canadian Foundation for Innovation, and the Social Science and Humanities Research Council of Canada. He is the education leadership team member for the McGill Collaborative for AI and Society.

In drug trials, lack of oversight of research ethics boards could put Canadian patients at risk

2025-10-29T14:56:12Z

Research ethics boards are supposed to ensure that, among other things, patients understand the nature of the research and have given informed consent. (Unsplash/Nappy)

New drug approvals by Health Canada are based on the results of clinical trials. But before clinical trials can go ahead, they need to be approved by ethics committees known as Research Ethics Boards (REBs).

Virtually all hospitals where research is conducted have REBs, as do universities and other institutions. The REBs are supposed to ensure that patients understand the nature of the research and have given informed consent, that the trials are conducted in an ethical way that minimizes any harm to them and that the investigators are competent to do the research.

Given the crucial role they play, it’s important that REBs are not influenced by factors like financial motives, conflicts of interest or the goals of drug companies. Without oversight, these factors may encroach on the decisions made by REBs in Canada.

REBs in Canada

All that Canada’s Food and Drug Regulations say about REBs is that they need to approve clinical trials.

The Tri-Council Policy Statement does lay out who needs to be on a REB and gives some details about how REBs should operate, but these regulations only apply to research that’s funded by the tri-council, comprising the Canadian Institutes of Health Research (CIHR), the National Science and Engineering Research Council and the Social Sciences and Humanities Research Council.

Canada has no accreditation or inspection system for REBs and no oversight mechanism for the way that they undertake their reviews. An article in the Journal of Law, Medicine and Ethics noted that: “Aside from identifying information on the REB and its chair, no further information about the REB or its review is required” by Health Canada.

There used to be a National Council on Ethics in Human Research. The organization largely provided education, but there was the possibility that it could have been transformed into a national accrediting and oversight body.

But in 2010, its funding from Health Canada and CIHR was pulled. In its place, the Canadian General Standards Board published the voluntary Canadian Standard for Research Ethics Oversight of Biomedical Clinical Trials, but this guidance was withdrawn in 2018 due to limited use and support for its revision.

The still existing Canadian Association of Research Ethics Boards operates as a forum for discussion and has no regulatory powers.

For-profit REBS

The absence of any standards and regulations is becoming increasingly problematic. At least 70 per cent of clinical trials are now being done in the community, outside of health-care institutions and their in-house REBs. In addition, drug companies, which sponsor the vast majority of clinical trials, want a quick turnaround in approval by REBs.

A report by the Law Commission of Canada described academic based REBs as:

“overburdened and … stretched to the breaking point … As the work becomes increasingly complicated with globalization, technology and commercialization, REBs are struggling to find committee chairs or even members.”

In response to the movement of trials into the community where they aren’t covered by institutional REBs, it’s reasonable to assume that the number of for-profit REBs has grown, although there are no definite estimates of their number. Drug companies pay these for-profit REBs a fee to review their trials.

Trudo Lemmens, professor and Scholl Chair in Health Law and Policy at the University of Toronto Faculty of Law, has argued that the credibility and integrity of the research review is compromised by the perception of a possible conflict-of-interest (COI) when commercial REBs approve a clinical trial.

If the REB turns down too many trials or demands costly changes to the research protocol, companies may be reluctant to continue to submit future research proposals to it. Although to date, there has not been any research to verify or refute this concern, Lemmens argues that the honesty of individual REB members is not enough to remedy this situation.

In early October 2025, the New York Times published an investigation of for-profit Institutional Review Boards in the United States. Institutional review boards are the American equivalent of REBs. The story focused on two companies that dominate the business: WCG and Advarra, the latter controlled by private equity.

According to the Times, both companies “have close corporate relationships with drugmakers. And both have become part of multipronged enterprises selling pharmaceutical companies a wide range of drug-testing services — blurring the line between the reviewer and the reviewed, introducing potential conflicts of interest that threaten the review boards’ mission.”

Several former Advarra employees told the Times that the company had imposed daily quotas on reviewing informed-consent forms for trial volunteers. Alana Levy, a former consent form development editor, said that falling short meant “you get a warning” but if you reviewed over a certain number you could get a bonus. Advarra refuted those allegations and said it “maintains strong safeguards and internal policies to ensure the independence of its Institutional Review Board.”

Advarra also operates in Canada and “supports more Canadian sites than any other partner, offering the broadest provincial coverage and experience in the industry.” On its website, it advertises the speed of its reviews with a turn-around time of four to five days for reviewing protocols and consent forms for trials taking place at multiple sites.

Oversight needed

When good ethical oversight is lacking, the patients in clinical trials may be put at risk. The results from those trials may be compromised, meaning that the information that doctors rely on to prescribe the drugs is unreliable, and their patients are getting suboptimal care.

Health Canada needs to step up and establish regulations for how REBs operate and have an inspection system to ensure that its regulations are being followed.

Alberta is the only jurisdiction in Canada without for-profit REBs. Among its other responsibilities, the Health Research Ethics Board of Alberta oversees ethics approval of research involving human subjects that is done in the community. Other provinces should follow the Alberta model.

Between 2022-2025, Joel Lexchin received payments for writing a brief for a legal firm on the role of promotion in generating prescriptions for opioids, for being on a panel about pharmacare and for co-writing an article for a peer-reviewed medical journal on semaglutide. He is a member of the Boards of Canadian Doctors for Medicare and the Canadian Health Coalition. He receives royalties from University of Toronto Press and James Lorimer & Co. Ltd. for books he has written. He has received funding from the Canadian Institutes of Health Research in the past.

New research reveals that almost half of Canadians believe in the paranormal — ghosts and all

2025-10-28T19:20:29Z

What would you say if you were told that paranormal activity exists? Well, nearly half of Canadians would agree.

What is the paranormal, exactly? It refers to phenomena that science cannot explain and are not part of a major religion in a particular society. In contrast, religious phenomena are part of an established doctrine. For example, in Canada, psychic abilities and Bigfoot or Sasquatch are considered paranormal, while angels and demons are associated with religion.

In the summer of 2025, we launched a survey of Canadian attitudes regarding paranormal beliefs in which participation was confidential. And for the first time in decades, we have nationally representative data on paranormal beliefs and encounters in Canada.

Although news outlets regularly publish stories about paranormal beliefs on Halloween, the results they discuss are usually based on convenience samples. Ours is the first study in 20 years to use randomly selected people from the Canadian population to ask these questions — meaning the results are representative.

And it turns out that almost one in two Canadians believe in at least one paranormal phenomenon, and one-quarter report encounters with spirits.

We asked about ghostly hauntings, alien visitations, psychic abilities, telekinesis, astrology and other unexplained phenomena. We also asked about cryptids — animals or creatures whose existence has been suggested but not (yet) proven by science — specific to Canada. They include creatures with roots in First Nations folklore like the large serpentine sea monster, the Cadborosaurus, off the B.C. coast and the Ogopogo in Lake Okanagan.

The believers, the skeptics and the in-between

Canada is one of the world’s most secular societies. Here, religion has little impact on the way people act or view the world.

How Canadians think about the paranormal, however, has been mostly unknown. It’s expensive to gather representative data in Canada and few social scientists think it’s important to study belief in the paranormal. The combination of these two factors has meant Canadian paranormal beliefs have gone unexamined for decades.

What we found is that Canadians have embraced the paranormal — to a point.

Almost half — 44 per cent — believe in at least one paranormal phenomenon. About one-third did not report belief in any paranormal phenomenon but did indicate neutrality about at least one. For example, several respondents did not believe in ghosts, but were on the fence about extraterrestrial visitations.

Many non-probability samples of Canadians have been surveyed over the last few years, but unlike ours, those results tell us little because they did not use random sampling to recruit respondents. This graph shows how many Canadians believe, are neutral or don’t believe in the existence of paranormal activity. (Sophia Dimitrakopoulos), CC BY-ND

Only about one-quarter said they did not believe in any of the 10 phenomena we asked about. The percentage of firm non-believers is similar to the 28 per cent figure in the United States and the United Kingdom Belief varied by specific phenomenon. People were most likely to believe in ghostly hauntings.

Respondents answered the authors’ survey on a granular lever, revealing whether they neither disagree nor agree, somewhat agree or strongly agree with whether each type of 10 paranormal phenomenon exists or not. (Sophia Dimitrakopoulos), CC BY-ND

Overall, it is more common for Canadians to believe in at least one paranormal phenomenon than to not believe in any.

Who is most likely to believe?

Patterns of belief vary somewhat by demographic group.

Women are more likely than men to believe in ghosts and psychics, reflecting how women have a higher probability of being open to phenomena with a spiritual dimension.

People with bachelor’s degrees or higher are less likely to believe in most paranormal phenomena. There are few racial or ethnic differences.

Interestingly, people aged 19-29 are less likely to believe in many paranormal phenomena than those aged 30-44 or 45-64. These findings suggest that young Canadians tend to opt out of any non-scientific belief system, whether religious or paranormal.

Few differences by region or language exist, though francophones are less likely to believe in Sasquatch than anglophones are.

Paranormal experiences in Canada

About one-quarter of Canadians claim to have heard, seen or felt a ghost or spirit. Some experiences were connected to religion, such as feeling the Christian Holy Spirit.

More often, experiences were associated with the death of a loved one and were personally meaningful. As one participant explained: “Soon after my mother’s death, I woke up suddenly and she was standing beside my bed. She smiled at me and faded away. I was comforted.”

Others reported spooky encounters associated with a place. A different participant wrote: “I was managing a motel and saw a ghostly man walking along the upper balcony. I asked the locals, and they said on the property that the motel was on, there was a house that burned down — and he lived in the house!”

Cryptid sightings are less common.

“I was operating a high-clearance sprayer, in a 1,300-acre field. I sat about 10 feet in the air in the cab on this machine,” one participant said. “I came around the corner of a bluff and saw a blurry, bipedal creature. It was furry, had a long snout and long arms, and in an instant turned into a moose. I have no idea to this day what that was.”

What our beliefs reveal

Our goal is not to prove or disprove any experience or belief, but to analyze what they mean for individuals and for Canada.

And to that end, our survey showed us that while many Canadians have replaced or supplemented religious belief with paranormal belief, most trust science. Belief in the paranormal or religion does not mean Canadians reject science, but rather that they believe some phenomena cannot yet be explained by science.

While the paranormal is fun — or creepy — to think about around Halloween, it is also part of the everyday belief system of many Canadians.

Tony Silva (as co-applicant) received funding from the Social Sciences and Humanities Research Council of Canada for the first survey wave of this project, which focused on attitudes about politics and decarbonisation. No grant or taxpayer funds were used for the second survey wave, which included questions about paranormal beliefs.

Emily Huddart received funding from the Social Sciences and Humanities Research Council of Canada to support an earlier wave of this project (with Tony Silva).

How to ensure youth, parents, educators and tech companies are on the same page on AI

2025-10-22T19:41:32Z

Artificial intelligence is now part of everyday life. It’s in our phones, schools and homes. For young people, AI shapes how they learn, connect and express themselves. But it also raises real concerns about privacy, fairness and control.

AI systems often promise personalization and convenience. But behind the scenes, they collect vast amounts of personal data, make predictions and influence behaviour, without clear rules or consent.

This is especially troubling for youth, who are often left out of conversations about how AI systems are built and governed.

The author’s guide on how to protect youth privacy in an AI world.

Concerns about privacy

My research team conducted national research and heard from youth aged 16 to 19 who use AI daily – on social media, in classrooms and in online games.

They told us they want the benefits of AI, but not at the cost of their privacy. While they value tailored content and smart recommendations, they feel uneasy about what happens to their data.

Many expressed concern about who owns their information, how it is used and whether they can ever take it back. They are frustrated by long privacy policies, hidden settings and the sense that you need to be a tech expert just to protect yourself.

As one participant said:

“I am mainly concerned about what data is being taken and how it is used. We often aren’t informed clearly.”

Uncomfortable sharing their data

Young people were the most uncomfortable group when it came to sharing personal data with AI. Even when they got something in return, like convenience or customization, they didn’t trust what would happen next. Many worried about being watched, tracked or categorized in ways they can’t see.

This goes beyond technical risks. It’s about how it feels to be constantly analyzed and predicted by systems you can’t question or understand.

AI doesn’t just collect data, it draws conclusions, shapes online experiences, and influences choices. That can feel like manipulation.

Parents and teachers are concerned

Adults (educators and parents) in our study shared similar concerns. They want better safeguards and stronger rules.

But many admitted they struggle to keep up with how fast AI is moving. They often don’t feel confident helping youth make smart choices about data and privacy.

Some saw this as a gap in digital education. Others pointed to the need for plain-language explanations and more transparency from the tech companies that build and deploy AI systems.

Professionals focus on tools, not people

The study found AI professionals approach these challenges differently. They think about privacy in technical terms such as encryption, data minimization and compliance.

While these are important, they don’t always align with what youth and educators care about: trust, control and the right to understand what’s going on.

Companies often see privacy as a trade-off for innovation. They value efficiency and performance and tend to trust technical solutions over user input. That can leave out key concerns from the people most affected, especially young users.

Power and control lie elsewhere

AI professionals, parents and educators influence how AI is used. But the biggest decisions happen elsewhere. Powerful tech companies design most digital platforms and decide what data is collected, how systems work and what choices users see.

Even when professional push for safer practices, they work within systems they did not build. Weak privacy laws and limited enforcement mean that control over data and design stays with a few companies.

This makes transparency and holding platforms accountable even more difficult.

What’s missing? A shared understanding

Right now, youth, parents, educators and tech companies are not on the same page. Young people want control, parents want protection and professionals want scalability.

These goals often clash, and without a shared vision, privacy rules are inconsistent, hard to enforce or simply ignored.

Our research shows that ethical AI governance can’t be solved by one group alone. We need to bring youth, families, educators and experts together to shape the future of AI.

The PEA-AI model

To guide this process, we developed a framework called PEA-AI: Privacy–Ethics Alignment in Artificial Intelligence. It helps identify where values collide and how to move forward. The model highlights four key tensions:

1. Control versus trust: Youth want autonomy. Developers want reliability. We need systems that support both.

2. Transparency versus perception: What counts as “clear” to experts often feels confusing to users.

3. Parental oversight versus youth voice: Policies must balance protection with respect for youth agency.

4. Education versus awareness gaps: We can’t expect youth to make informed choices without better tools and support.

What can be done?

Our research points to six practical steps:

1. Simplify consent. Use short, visual, plain-language forms. Let youth update settings regularly.

2. Design for privacy. Minimize data collection. Make dashboards that show users what’s being stored.

3. Explain the systems. Provide clear, non-technical explanations of how AI works, especially when used in schools.

4. Hold systems accountable. Run audits, allow feedback and create ways for users to report harm.

5. Teach privacy. Bring AI literacy into classrooms. Train teachers and involve parents.

6. Share power. Include youth in tech policy decisions. Build systems with them, not just for them.

AI can be a powerful tool for learning and connection, but it must be built with care. Right now, our research suggests young people don’t feel in control of how AI sees them, uses their data or shapes their world.

Ethical AI starts with listening. If we want digital systems to be fair, safe and trusted, we must give youth a seat at the table and treat their voices as essential, not optional.

Ajay Shrestha receives funding from the Office of the Privacy Commissioner of Canada (OPC); the views expressed herein are those of the authors and do not necessarily reflect those of the OPC.

A digital twin could help Canada beat wildfires, fix commutes and save tax dollars

2025-10-15T17:29:35Z

Canada is facing larger wildfires, rising flood risks and worsening traffic congestion. The federal government’s infrastructure plan budgets at least $180 billion over 12 years, yet insured disaster losses hit a record $8.5 billion in 2024.

Despite these massive investments, too often problems are only discovered after construction begins. One way to address this is to model risks and impacts before they occur using a digital replica that mirrors how real systems work.

A “digital twin” — essentially a live virtual model of roads, transit, energy, water and public buildings — would let policymakers and planners test ideas and spot risks ahead of time. It blends maps and 3D models with data (some live, some updated regularly), so policymakers and planners can run “what-if” scenarios.

For example, policymakers could use a digital twin to see how a lane closure, new bus route or wildfire evacuation order might ripple through a city before making a decision. Singapore already uses this approach to test planning and emergency responses and its documented efficiency gains are clear.

As researchers, we see a national, federated digital twin improving Canada’s resilience and efficiency in three practical ways.

Benefit #1: Safer wildfire evacuations

Canada’s 2023 wildfire season was the worst on record, with more than 18 million hectares burned, and 2025 has already been called the second-worst on record.

When fires move fast, evacuation routes can become jammed and communication can break down. During the 2016 Fort McMurray wildfire, for instance, residents received “mixed messages” about the threat and proximity of the wildfire. Thousands of people ended up jamming Highway 63, the sole road in and out of the city.

Similarly, during Yellowknife’s 2023 evacuation, an after-action review found there was a lack of clear and transparent communication to the public about an evacuation plan, leading to “significant confusion and stress.”

A national digital twin could help emergency teams rehearse evacuations in advance. They could test detours, traffic signal plans, one-way controls, signage and reception-centre capacity; check if ambulances can reach hospitals when smoke closes a route; and push clear routes to navigation apps in real time.

Benefit #2: Faster, more reliable commutes

Traffic congestion and transit delays cost Canadians time, productivity and peace of mind. We all know what it’s like when a construction project snarls traffic or a crowded station slows trains.

A 2024 report from the Canadian Centre for Economic Analysis estimated that congestion cost Ontario $56.4 billion in total economic and social impacts. Of that, about $43.6 billion was linked to reduced quality of life, including stress, health impacts and time lost to delays.

A digital twin could help. With this technology, transit agencies could test bus-only lanes, signal timing, platform-crowding fixes and construction plans before rolling them out.

Vancouver International Airport has already built a real-time digital twin to optimize passenger flows. The same principles can also be applied to transit hubs and busy corridors, helping cities identify problems early, reduce disruption and move people more efficiently.

Benefit #3: Better use of tax dollars

Cost overruns and rework continue to drain public budgets across Canada. Major infrastructure projects frequently exceed their initial pricetags, like the Trans Mountain pipeline expansion, which is now projected to cost nearly $34 billion — almost six times the original $5.3 billion estimate.

Montréal’s Réseau express metropolitain light-rail project has faced multiple cost increases as wells, rising from an initial estimate of $6.3 billion to more than $7.9 billion as of 2023.

Digital twins can reduce these losses by identifying design conflicts early, comparing options side-by-side and improving transparency with the public.

Evidence suggests the savings can be substantial. A technical report from the National Research Council of Canada found that using digital design tools to resolve design conflicts early saved roughly 20 per cent of a project’s contract value.

The potential returns are equally clear abroad. The U.K. government estimates that applying digital twins to network management could deliver 856 million pounds in benefits over 10 years.

Canada is already testing these possibilities. Ontario’s $5 million digital twin pilot is exploring how they can be used to deliver hospitals, highways and transit projects on time and on budget.

Similarly, the federal government is exploring using a digital twin to improve infrastructure maintenance and planning. Public Services and Procurement Canada has issued a Request for Information on a digital twin platform for its building portfolio.

From scattered projects to a national framework

Canada already has a strong foundation to build on for a national digital twin. Many Canadian cities already publish detailed base spatial data, such as Toronto’s 3D massing models and Vancouver’s public LiDAR data that captures its urban form in high resolution.

Canadian universities are already leading the way. Researchers at Carleton University have been the first to model a digital twin at a national scale, and plan to release the project’s code as an open-source project and the platform for free.

Infrastructure Ontario and Toronto Metropolitan University have signed a two-year partnership to apply digital-twin technology to modernize provincial infrastructure planning. Meanwhile, four other Canadian universities are involved in a project to explore how these tools can improve development approvals and regulatory decision-making.

The challenge is not to start from scratch, but to connect these existing initiatives under a coherent national framework.

This means agreeing on a few shared rules: common formats so maps and assets line up, clear privacy and security standards that prohibit personal tracking (only anonymous or aggregated data) and a small federal team to maintain standards and allow the different systems to work together.

Transparency about how the digital twin models work will be essential. The government should publish the methods and test results online for communities, journalists and independent experts to check. Routine audits and a quick way to fix mistakes should also be added.

A practical first step is to focus on projects that address urgent, tangible issues, namely wildfire evacuation routes and commute reliability. Early successes in these areas would demonstrate value quickly while proving the model’s effectiveness.

Learning from global leaders

Canada does not need to invent its own rule book. It can adopt existing frameworks like the U.K.’s plain-English Gemini Principles and information-management playbook, which focuses on public benefit, openness and safety.

Singapore, the U.K. and the European Union have all developed, implemented and tested digital twin programs, showing how to set standards, protect privacy and deliver public benefits.

If Canada borrows their templates and lessons, it can move faster and at a lower cost. It will be able to link early adopters, focus on high-impact uses like wildfire evacuations and commute reliability, publish results for review and then expand.

By doing so, Canada would shift from fragmented projects to a national digital twin that strengthens resilience, protects privacy and improves everyday life.

Dr. Ehsan Noroozinejad has received funding from both national and international organizations to support research addressing housing and climate crises. His most recent funding for integrated housing and climate policy comes from the APPI. He has also been involved in securing funding from NSERC and Mitacs. He is also affiliated with Western Sydney University.

Professor T.Y. Yang secures funding from national and international organizations to develop innovative solutions for housing and climate crises, with a focus on modern methods of construction.

Why industry-standard labels for AI in music could change how we listen

2025-10-13T10:09:24Z

Earlier this year, a band called The Velvet Sundown racked up hundreds of thousands of streams on Spotify with retro-pop tracks, generating a million monthly listeners on Spotify.

But the band wasn’t real. Every song, image, and even its back story, had been generated by someone using generative AI.

For some, it was a clever experiment. For others, it revealed a troubling lack of transparency in music creation, even though the band’s Spotify descriptor was later updated to acknowledge it is composed with AI.

In September 2025, Spotify announced it is “helping develop and will support the new industry standard for AI disclosures in music credits developed through DDEX.” DDEX is a not-for-profit membership organization focused on the creation of digital music value chain standards.

The company also says it’s focusing work on improved enforcement of impersonation violations and a new spam-filtering system, and that updates are “the latest in a series of changes we’re making to support a more trustworthy music ecosystem for artists, for rights-holders and for listeners.”

As AI becomes more embedded in music creation, the challenge is balancing its legitimate creative use with the ethical and economic pressures it introduces. Disclosure is essential not just for accountability, but to give listeners transparent and user-friendly choices in the artists they support.

A patchwork of policies

The music industry’s response to AI has so far been a mix of ad hoc enforcement as platforms grapple with how to manage emerging uses and expectations of AI in music.

Apple Music took aim at impersonation when it pulled the viral track “Heart on My Sleeve” featuring AI-cloned vocals of Drake and The Weeknd. The removal was prompted by a copyright complaint reflecting concerns over misuse of artists’ likeness and voice.

CBC News covers AI-generated band The Velvet Sundown.

The indie-facing song promotion platform SubmitHub has introduced measures to combate AI-generated spam. Artists must declare if AI played “a major role” in a track. The platform also has an “AI Song Checker” so playlist curators can scan files to detect AI use.

Spotify’s announcement adds another dimension to these efforts. By focusing on disclosure, it recognizes that artists use AI in many different ways across music creation and production. Rather than banning these practices, it opens the door to an AI labelling system that makes them more transparent.

Labelling creative content

Content labelling has long been used to help audiences make informed choices about their media consumption. Movies, TV and music come with parental advisories, for example.

Digital music files also include embedded information tags called metadata, which include details like genre, tempo and contributing artists that platforms use to categorize songs, calculate royalty payments and to suggest new songs to listeners.

Canada has relied on labelling for decades to strengthen its domestic music industry. The MAPL system requires radio stations to play a minimum percentage of Canadian music, using a set of criteria to determine whether a song qualifies as Canadian content based on music, artist, production and lyrics.

As more algorithmically generated AI music appears on streaming platforms, an AI disclosure label would give listeners a way to discover music that matches their preferences, whether they’re curious about AI collaboration or drawn to more traditional human-crafted approaches.

What could AI music labels address?

A disclosure standard will make AI music labelling possible. The next step is cultural: deciding how much information should be shared with listeners, and in what form.

According to Spotify, artists and rights-holders will be asked to specify where and how AI contributed to a track. For example, whether it was used for vocals, instrumentation or post-production work such as mixing or mastering.

For artists, these details better reflect how AI tools fit into a long tradition of creative use of new technologies. After all, the synthesizer, drum machines and samplers — even the electric guitar — were all once controversial.

But AI disclosure shouldn’t give streaming platforms a free pass to flood catalogues with algorithmically generated content. The point should also be to provide information to listeners to help them make more informed choices about what kind of music they want to support.

Information about AI use should be easy to see and quickly find. But on Spotify’s Velvet Sundown profile, for example, this is dubious: listeners have to dig down to actually read the band’s descriptor.

AI and creative tensions in music

AI in music raises pressing issues, including around labour and compensation, industry power dynamics, as well as licensing and rights.

One study commissioned by the International Confederation of Societies of Authors and Composers has said that Gen AI outputs could put 24 per cent of music creators’ revenues at risk by 2028, at a time when many musician careers are already vulnerable to high costs of living and an unpredictable and unstable streaming music economy.

The most popular AI music platforms are controlled by major tech companies. Will AI further concentrate creative power, or are there tools that might cut production costs and become widely used by independent artists? Will artists be compensated if their labels are involved in deals for artists’ music to train AI platforms?

The cultural perception around musicians having their music train AI platforms or in using AI tools in music production is also a site of creative tension.

Enabling listener choice

Turning a disclosure standard into something visible — such as an intuitive label or icon that allows users to go deeper to show how AI was used — would let listeners see at a glance how human and algorithmic contributions combine in a track.

Embedded in the digital song file, it could also help fans and arts organizations discover and support music based on the kind of creativity behind it.

Ultimately, it’s about giving listeners a choice. A clear, well-designed labelling system could help audiences understand the many ways AI now shapes music, from subtle production tools to fully synthetic vocals.

Need for transparency

As influence of AI in music creation continues to expand, listeners deserve to know how the sounds they love are made — and artists deserve the chance to explain it.

Easy-to-understand AI music labels would turn disclosure into something beyond compliance: it might also invite listeners to think more deeply about the creative process behind the music they love.

Gordon A. Gow receives funding from the Social Sciences and Humanities Research Council (SSHRC).

Brian Fauteux does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

OpenAI’s newly launched Sora 2 makes AI’s environmental impact impossible to ignore

2025-10-09T21:03:01Z

OpenAI’s recent rollout of its new video generator Sora 2 marks a watershed moment in AI. Its ability to generate minutes of hyper-realistic footage from a few lines of text is astonishing, and has raised immediate concerns about truth in politics and journalism.

But Sora 2 is rolling out slowly because of its enormous computational demands, which point to an equally pressing question about generative AI itself: What are its true environmental costs? Will video generation make them much worse?

The recent launch of the Stargate Project — a US$500 billion joint venture between OpenAI, Oracle, SoftBank and MGX — to build massive AI data centres in the United States underscores what’s at stake. As companies race to expand computing capacity on this scale, AI’s energy use is set to soar.

The debate over AI’s environment impact remains one of the most fraught in tech policy. Depending on what we read, AI is either an ecological crisis in the making or a rounding error in global energy use. As AI moves rapidly into video, clarity on its footprint is more urgent than ever.

OpenAI showcases Sora 2’s capabilities.

Two competing narratives

From one perspective, AI is rapidly becoming a major strain on the world’s energy and water systems.

Alex de Vries-Gao, a researcher who has long tracked the electricity use of bitcoin mining, noted in mid-2025 that AI was on track to surpass it. He estimated that AI already accounted for about 20 per cent of global data-center power consumption; this is likely to double by year’s end.

According to the International Energy Agency, data centres used up to 1.5 per cent of global electricity consumption last year, with consumption growing four times faster than total global demand. The IEA predicts that data centres will more than double their use by 2030, with AI processing the leading driver of growth.

Research cited by MIT’s Technology Review concurs, estimating that by 2028, AI’s power draw could exceed “all electricity currently used by US data centers” — enough to power 22 per cent of U.S. households each year.

‘Huge’ quantities

AI’s water use is also striking. Data centres rely on ultra-pure water to keep servers cool and free of impurities. Researchers estimated that training GPT-3 would have used up 700,000 litres of freshwater at Microsoft’s American facilities. They predict that global AI demand could reach four to six billion cubic metres annually by 2027.

Hardware turnover adds further strain. A 2023 study found that chip fabrication requires “huge quantities” of ultra-pure water, energy-intensive chemical processes and rare minerals such as cobalt and tantalum. Manufacturing the high-end graphics processing units — the engines that drive AI boom — has a much larger carbon footprint than most consumer electronics.

Generating an image uses the electricity of a microwave running for five seconds, while making a five-second video clip takes up as much as a microwave running for over an hour.

The next leap from text and image to high-definition video could dramatically increase AI’s impact. Early testing bears this out — finding that energy use for text-to-video models quadruples when video length doubles.

The case for perspective

Others see the alarm as overstated. Analysts at the Center for Data Innovation, a technology and policy think tank, argue that many estimates about AI energy use rely on faulty extrapolations. GPU hardware is becoming more efficient each year, and much of the electricity in new data centres will come from renewables.

Recent benchmarking puts AI’s footprint in context. Producing a typical chatbot Q&A consumes about 2.9 watt-hours (Wh) — roughly 10 times a Google search. Google recently claimed that a typical Gemini prompt uses only 0.24 Wh and 0.25 mL of water, though independent experts note those numbers omit indirect energy and water used in power generation.

Context is key. An hour of high-definition video streaming on Netflix uses roughly 100 times more energy than generating a text response. An AI query’s footprint is tiny, yet data centres now process billions daily, and more demanding video queries are on the horizon.

Jevons paradox

It helps to distinguish between training and use of AI. Training frontier models such as GPT-4 or Claude Opus 3 required thousands of graphics chips running for months, consuming gigawatt-hours of power.

Using a model takes up a tiny amount of energy per query, but this happens billions of times a day. Eventually, energy from using AI will likely surpass training.

The least visible cost may come from hardware production. Each new generation of chips demands new fabrication lines, heavy mineral inputs and advanced cooling. Italian economist Marcello Ruberti observes that “each upgrade cycle effectively resets the carbon clock” as fabs rebuild highly purified equipment from scratch.

And even if AI models become more efficient, total energy keeps climbing. In economics, this is known as the Jevons paradox: in 19th-century Britain, the consumption of coal increased as the cost of extracting it decreased. As AI researchers have noted, as costs per-query fall, developers are incentivized to find new ways to embed AI into every product. The result is more data centres, chips and total resource use.

A problem of scale

Is AI an ecological menace or a manageable risk? The truth lies somewhere in between.

A single prompt uses negligible energy, but the systems enabling it — vast data centres, constant chip manufacturing, round-the-clock cooling — are reshaping global energy and water patterns.

The International Energy Agency’s latest outlook projects that data-centre power demand could reach 1,400 terawatt-hours by 2030. This is the equivalent of adding several mid-sized countries to the world’s grid. AI will count for a quarter of that growth.

Transparency is vital

Many of the figures circulating about AI energy use are unreliable because AI firms disclose so little. The limited data they release often employ inconsistent metrics or offset accounting that obscures real impacts.

One obvious fix would be to mandate disclosure rules: standardized, location-based reporting of the energy and water used to train and operate models. Europe’s Artificial Intelligence Act requires developers of “high-impact” systems to document computation and energy use.

Similar measures elsewhere could guide where new data centres are built, favouring regions with abundant renewables and water — this could encourage longer hardware lifecycles instead of annual chip refreshes.

Balancing creativity and cost

Generative AI can help unlock extraordinary creativity and provide real utility. But each “free” image, paragraph or video has hidden material and energy costs.

Acknowledging those costs doesn’t mean we need to halt innovation. It means we should demand transparency about how great the environmental cost is, and who pays it, in order to address AI’s environmental impacts.

As Sora 2 begins to fill social feeds with highly realistic visuals, the question won’t be whether AI uses more energy than Netflix, but whether we can expand our digital infrastructure responsibly enough to make room for both.

Robert Diab does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

What AI-generated Tilly Norwood reveals about digital culture, ethics and the responsibilities of creators

2025-10-08T14:04:22Z

Imagine an actor who never ages, never walks off set or demands a higher salary.

That’s the promise behind Tilly Norwood, a fully AI-generated “actress” currently being courted by Hollywood’s top talent agencies. Her synthetic presence has ignited a media firestorm, denounced as an existential threat to human performers by some and hailed as a breakthrough in digital creativity by others.

But beneath the headlines lies a deeper tension. The binaries used to debate Norwood — human versus machine, threat versus opportunity, good versus bad — flatten complex questions of art, justice and creative power into soundbites.

The question isn’t whether the future will be synthetic; it already is. Our challenge now is to ensure that it is also meaningfully human.

All agree Tilly isn’t human

Ironically, at the centre of this polarizing debate is a rare moment of agreement: all sides acknowledge that Tilly is not human.

Her creator, Eline Van der Velden, the CEO of AI production company Particle6, insists that Norwood was never meant to replace a real actor. Critics agree, albeit in protest. SAG-AFTRA, the union representing actors in the U.S., responded with:

“It’s a character generated by a computer program that was trained on the work of countless professional performers — without permission or compensation. It has no life experience to draw from, no emotion, and from what we’ve seen, audiences aren’t interested in watching computer-generated content untethered from the human experience.”

Their position is rooted in recent history: In 2023, actors went on strike over AI. The resulting agreement secured protections around consent and compensation.

So if both sides insist Tilly isn’t human, the controversy, then, isn’t just about what Tilly is, it’s about what she represents.

Complexity as a starting point

Norwood represents more than novelty. She’s emblematic of a larger reckoning with how rapidly artificial intelligence is reshaping our lives and the creative sector. The velocity of change is dizzying, and now the question is how do we shape the hybrid world we’ve already entered?

It can feel disorienting trying to parse ethics, rights and responsibilities while being bombarded by newness. Especially when that “newness” comes in a form that unnerves us: a near-human likeness that triggers long-standing cultural discomfort.

Indeed, Norwood may be a textbook case of the “uncanny valley,” a term coined by Japanese roboticist Masahiro Mori to describe the unease people feel when something looks almost human, but not quite.

But if all sides agree that Tilly isn’t human, what happens when audiences still feel something real while watching her on screen? If emotional resonance and storytelling are considered uniquely human traits, maybe the threat posed by synthetic actors has been overstated. On the other hand, who hasn’t teared up in a Pixar film? A character doesn’t have to feel emotion to evoke it.

Still, the public conversation remains polarized. As my colleague Owais Lightwala, assistant professor in the School of Performance at Toronto Metropolitan University, puts it: “The conversation around AI right now is so binary that it limits our capacity for real thinking. What we need is to be obsessed with complexity.”

Synthetic actors aren’t inherently villains or saviours, Lightwala tells me, they’re a tool, a new medium. The challenge lies in how we build the infrastructures around them, such as rights, ownership and distribution.

He points out that while some celebrities see synthetic actors as job threats, most actors already struggle for consistent work. “We ask the one per cent how they feel about losing power, but what about the 99 per cent who never had access to that power in the first place?”

Too often missing from this debate is what these tools might make possible for the creators we rarely hear from. The current media landscape is already deeply inequitable. As Lightwala notes, most people never get the chance to realize their creative potential — not for lack of talent, but due to barriers like access, capital, mentorship and time.

Now, some of those barriers might finally lower. With AI tools, more people may get the opportunity to create.

Of course, that doesn’t mean AI will automatically democratize creativity. While tools are more available, attention and influence remain scarce.

Sarah Watling, co-founder and CEO of JaLi Research, a Toronto-based AI facial animation company, offers a more cautionary perspective. She argues that as AI becomes more common, we risk treating it like a utility, essential yet invisible.

In her view, the inevitable AI economy won’t be a creator economy, it will be a utility commodity. And “when things become utilities,” she warns, “they usually become monopolized.”

Where do we go from here?

We need to pivot away from reactionary fear narratives, like Lightwala suggests.

Instead of shutting down innovation, we need to continue to experiment. We need to use this moment, when public attention is focused on the rights of actors and the shape of culture, to rethink what was already broken in the industry and allow space for new creative modalities to emerge.

Platforms and studios must take the lead in setting transparent, fair policies for how synthetic content is developed, attributed and distributed. In parallel, we need to push creative institutions, unions and agencies to collaborate in the co-design of ethical and contractual guardrails now, before precedents get set in stone, putting consent, fair attribution and compensation at the centre.

And creators, for their part, must use these tools not just to replicate what came before, but to imagine what hasn’t been possible until now. That responsibility is as much creative as it is technical.

The future will be synthetic. Our task now is to build pathways, train talent, fuel imagination, and have nuanced, if difficult, conversations. Because while technology shapes what’s possible, creators and storytellers have the power to shape what matters.

Ramona Pringle does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

The H-1B visa fee hike in the United States opens a policy window for Canada

2025-10-07T13:43:37Z

The MaRS urban innovation hub building in Toronto. Canada may benefit from the American H-1B visa fee increase by attracting highly skilled tech workers and others from abroad to Canada instead of the United States. (WikiMedia), CC BY

The United States government recently announced a US$100,000 H-1B visa fee on new applications, which will affect highly educated workers from abroad who are seeking jobs in the U.S. This policy could have ripple effects for Canada by reducing the emigration of Canadians going to work in the U.S. — and by attracting a talented workforce to the country.

The H-1B visa program was created in 1990 for applicants with at least a bachelor’s degree or higher to work in the U.S. The current annual statutory cap is 65,000 visas, with 20,000 additional visas for professionals from abroad who graduate with a master’s or doctorate from an American institution of higher learning.

The recent announcement regarding the fee increase has astounded tech companies that have long relied on the visa to employ foreign workers in the U.S. Since 2012, about 60 per cent of H-1B workers approved each year have held a tech-related job. Tech companies have been pushing U.S. Congress to expand the visa program due to the high demand and competition for the H-1B.

Instead, this massive increase in fees will make it much more expensive for firms to hire highly educated and skilled immigrants.

The impact on Canadians

Approximately 828,000 Canadian-born immigrants lived in the U.S. as of 2023, many of whom moved to the country via employment channels. Canadians made up one per cent of the total H-1B applications in 2019, and the new H-1B visa fee could reduce the number of Canadians moving to the U.S. for work.

This is especially true in the tech sector, as noted by Prime Minister Mark Carney in his recent remarks at the Council of Foreign Relations in New York:

“We are a leading developer of AI. And our research universities are some of the biggest producers in volume of AI, computing and quantum talent in the world. Unfortunately, most of them go to the United States. I understand you’re changing your visa policy, I hear, so going to hang onto a few of those.”

But another possibility is that American businesses could shift towards using the TN visa — an American non-immigrant visa for citizens of Canada and Mexico to work in specific professional-level jobs — to hire more Canadian workers. Canadians are eligible for the work permit under the Canada-US-Mexico-trade agreement.

These companies could then bypass paying the new H-1B visa fee while still hiring Canadian talent.

Prime Minister Mark Carney’s full remarks to the Council of Foreign relations. (Reuters)

Competition for global talent

Various countries, including Canada, are competing to attract and retain global talent. For many highly educated people from abroad seeking work in the U.S., especially recent international graduates of American universities, the new visa fee might result in fewer employment opportunities. As they start to look elsewhere, Canada could be an attractive destination if immigration pathways can be provided in a timely fashion.

Research has also shown that when faced with restrictions on immigration policies to hire skilled immigrants, U.S.-based multinational companies have responded by decreasing the number of jobs they offer in the U.S. and by increasing foreign affiliate employment, particularly in India, China and Canada.

So Canada should be proactive in working with these companies as they plan alternate pathways to retain their workforce.

This sudden and drastic change in the H-1B visa fee by the Donald Trump government presents a window of opportunity for Canadian policymakers to react quickly and offer pathways to recruit such foreign talent. The Canadian government seems to be paying attention. Carney told a recent news conference in London:

“Not as many of those people are going to get visas to the United States. And these are people with lots of skills that are enterprising, and they’re willing to move to work …. So it’s an opportunity for Canada, and we’re going to take that into account. And we’ll have a clear offering on that.”

Crises can create opportunities

A policy window opens when there is the right combination of recognizing a problem and providing a feasible policy solution while there is a favourable political climate. This allows policymakers to link the problem to a solution and advocate for change.

In the current environment, policy officials inside and outside of government can provide ideas on creating targeted policies and pathways to recruit talented workers to Canada.

An example of such a targeted initiative was seen in 2023, when the Canadian government, while announcing its Tech Talent Strategy, introduced a program that allowed H-1B visa holders to apply to receive an open three-year work permit in Canada.

It became clear that Canada was regarded as a popular alternative when applications closed within 24 hours after the maximum number of 10,000 applications was reached.

Richa Shivakoti receives funding from the Social Science and Humanities Research Council.

Anna Triandafyllidou receives funding for research related to high-skilled migration and its governance from the Tri-Council Agency and the Social Sciences and Humanities Research Council of Canada, as well as Horizon Europe (Link4Skills research project).

Smartphones manipulate our emotions and trigger our reflexes — no wonder we’re addicted

2025-10-07T13:09:58Z

The frequency and length of daily phone use continues to rise, especially among young people. It’s a global concern, driving recent decisions to ban phones in schools in Canada, the United States and elsewhere.

Social media, gaming, streaming and interacting with AI chatbots all contribute to this pull on our attention. But we need to look at the phones themselves to get the bigger picture.

As I argue in my newly published book, Needy Media: How Tech Gets Personal, our phones — and more recently, our watches — have become animated beings in our lives. These devices can build bonds with us by recognizing our presence and reacting to our bodies.

Packed with a growing range of technical features that target our sensory and psychological soft spots, smartphones create comforting ties that keep us picking them up. The emotional cues designed into these objects and interfaces imply that they need our attention, while in actuality, the devices are soaking up our data.

A responsive presence

Face recognition, geolocation, touchscreens, vibration, sound alerts and audio and motion sensing all play their part in catching our attention and responding to our actions. Separately, these may not create a strong emotional attachment, but collectively they situate the phone as a uniquely intimate, sensitive and knowing presence in our lives.

Take facial recognition locks, for example. Convenient for quick access, a smartphone will light up and unlock with a glance when it encounters a known and trusted face. When introducing Face ID in 2017, Apple claimed: “Do it up anyway you do it, Face ID learns your face. It learns who you are.” This implies a deeper user-device connection, like the one we have with folks we know when we spot them crossing our path.

Some devices have repurposed the hand wave — a typical gesture of friendship — into a feature that triggers the camera to take a photo.

Geolocation converts networking signals into a dot on a map, and we see that dot as us — not our phone — just as we may see the dots of our friends’ phones on the map as them.

Phantom vibrations

Sensory cues play a strong role. Touchscreens allow the phone’s interface to react subtly, like edge lighting and rubberbanding, to mimic the pliability of skin.

Vibration and sound alerts make us highly sensitive to the smallest movement or sound from the device. This produces conditions like phantom vibration syndrome, where we imagine that the device requires our attention, even when it doesn’t.

Audio and motion sensing, on the other hand, allows the device to react to us almost instantly, as when it lowers its ringing on an incoming call when we grab its body.

Phones are constant companions as we move through our days. (Muradi/Unsplash), CC BY

Roots and origins

Most of these features were developed decades ago for other uses. GPS was created by the U.S. military in the early 1970s, then was adopted by hikers and sailors to both navigate and to allow others to locate them if necessary.

Vibration alerts were created for pagers in the late 1970s for professionals — from hospital staff to travelling salespeople — to notify them of an important phone call.

Sound alerts became more widespread with Tamagotchi and other 1990s digital pets. Those toys are especially significant when discussing today’s psychological dependency on portable devices.

Through their beeping cries for attention, Tamagotchi trained millions of school-age millennials to build emotional attachments to virtual handheld companions needing care and nurturing. Not surprisingly, these toys were banned in many schools for their tendency to disrupt classes and distract students.

Indiscriminate tracking

Phones have become an essential part of who we are and how we behave. But there’s also an issue of privacy around our most intimate actions and behaviours. Sensors keep sensing, measuring sounds, movements and proximity.

There is the risk that our dependency will intensify as phones learn things about us that have, until recently, been off limits.

Sleep is a good example. Audio and motion sensing allows the device to get a reasonable picture of when and how we sleep, often collecting and sharing biometric data through pre-loaded health and wellness apps.

Another example is more sophisticated facial recognition, that will not only be able to recognize a face, but also analyze expressions to determine alertness or mood.

All of this collected data may have profound consequences, making our bodily behaviour, our off-line interactions with others and our emotional fragility a regular part of the data profiles used to leverage our lives for corporate profit.

Managing dependency

Short of powering off or walking away, what can we do to manage this dependency? We can access device settings and activate only those features we truly require, adjusting them now and again as our habits and lifestyles change.

Turning on geolocation only when we need navigation support, for example, increases privacy and helps break the belief that a phone and a user are an inseparable pair. Limiting sound and haptic alerts can gain us some independence, while opting for a passcode over facial recognition locks reminds us the device is a machine and not a friend. This may also make it harder for others to access the device.

So-called “dumb phones” limit what a user can do with their devices, though that’s a tough sell when 24/7 connectivity is becoming an expectation.

Manufacturers can do their part by placing more invasive device settings in the “off” position in the factory and being more transparent about their potential uses and data liabilities. That’s not likely to happen, however, without stronger government regulation that puts users and their data first.

In the meantime, at a minimum, we should broaden our public discussions of dependency beyond social media, gaming and artificial intelligence to acknowledge how phones, in themselves, can capture our attention and cultivate our loyalty.

Stephen Monteiro does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

More than a quarter of Canadian teens have experienced sexual violence online

2025-10-06T13:24:44Z

Technology-facilitated sexual violence includes harmful practices such as sexual name-calling, rumour spreading, non-consensual distribution of nudes, and other forms of sexual harassment. (imgix/Unsplash), CC BY

Law enforcement agencies across Canada are sounding the alarm over a rise in sexual extortion (“sextortion”) against young people.

The problem goes far beyond sextortion, as this is only one form of many variations of online sexual harms that target youth today. Teenagers in Canada can be victims of sexual catfishing, AI-generated sexual deepfakes and violent extremism.

Some high-profile sextortion incidents include the deaths by suicide of Rehtaeh Parsons, Amanda Todd, Daniel Lints and a boy in British Columbia.

The scale of the problem

Technology-facilitated sexual violence (TFSV) includes harmful practices such as sexual name-calling and rumour spreading, non-consensual distribution of intimate images (nudes) and other forms of sexual harassment.

Our research team recently conducted a survey with more than 1,000 teens aged 13 to 18 across Canada to learn about youths’ experiences with TFSV.

Our findings underscore how widespread these harms are: more than a quarter of the teens (28 per cent) reported experiencing at least one form of TFSV.

In addition to so many youth experiencing TFSV, almost half (47 per cent) said that TFSV had happened to someone they knew. The most common forms of TFSV reported in our survey were receiving unwanted sexual images (15 per cent), encountering unwanted porn (13 per cent) and being sexually harassed online (11 per cent).

Online platforms

We also asked the teens which social media sites and online gaming services had the most sexual harassment. The platform they mentioned most often was Snapchat, followed by TikTok and Instagram. Snapchat has been known for its potential risks to youth and privacy concerns.

Girls experienced TFSV at a higher rate (32 per cent) than boys (23 per cent), which is consistent with research from Statistics Canada.

Teens who said they were neurodivergent or had a learning disability were more likely to be subjected to TFSV (39 per cent). TFSV was also higher among teens with a mental health condition (40 per cent).

These findings are consistent with previous research that showed higher victimization rates among people with disabilities.

A small but significant number of teens (seven per cent) reported committing at least one form of TFSV. This was more common among boys (nine per cent) than girls (six per cent).

Insufficient support

Parents and guardians were the most relevant source of support for teens who had been subjected to TFSV. Nearly half (44 per cent) of the teens turned to them, and most of these teens found them helpful.

The teens were much less likely to seek support from institutions. Only about one in 10 (12 per cent) told someone at their school, with only seven per cent telling the police. Unfortunately, these numbers are consistent with other statistics, as most people do not report sexual violence to the police.

Young people showed little confidence in the reporting tools and moderation systems of social media platforms. As little as five per cent of the teens had used these to report sexually harmful materials. Almost one in three teens (29 per cent) thought that the digital platforms should do a better job supporting them.

This finding is important to consider as social media companies are dropping content moderation, making their platforms possibly more hazardous for youth.

Teens’ misconceptions

Most teens (90 per cent and up) knew that several forms of TFSV were illegal in Canada. However, they were less certain when asked if it was legal to create a fake sexual video of someone. This is unsurprising: legal views of sexual deepfakes vary by province. Some allow civil action, while others treat it as child pornography.

The teens’ knowledge of the law was incomplete in other areas. Almost two-thirds (61 per cent) thought that sending a nude picture of themselves to other youth was illegal. This is not true. Minors can share sexual images with each other as long as they are consensual and kept private between them; that most teens don’t know this is troubling.

Sexting and sharing nudes is a common form of sexual expression among teens. In our survey, teens who though that nude image sharing was illegal were less likely to seek help with TFSV.

Some teens (26 per cent) thought that taking a nude picture of themselves was illegal. This is also incorrect.

These misconceptions matter, as young people need to be informed about their legal rights to sexual expression. Proper education will prevent shame, fear and other barriers to seeking support when someone is distributing their images against their will or coercing them into harmful practices.

The use of Snapchat by teens has raised concerns about its potential risks and privacy issues. (Souvik Banerjee/Unsplash), CC BY

An urgent issue

Social media and other forms of digital communication are central to young people’s lives, which means that addressing TFSV is an urgent issue. While the federal government and some provincial governments have taken steps or proposed legislation aimed at protecting youth, some responses have been proven to be unrealistic and ineffective.

Governments — and tech companies in particular — need to do more to prevent TFSV and support youth who experience it.

Schools can also take action to help youth. However, there is considerable variation in the TFSV responses and interventions within educational curricula, policies and legislation across the provinces and territories. This means that even though TFSV is a common problem, most parents, teachers, police and frontline workers lack the resources and strategies needed to respond effectively and promptly.

Our findings highlight the impact of these shortcomings on teens, as many youth in our survey did not receive help for TFSV, even when they sought it out. In many instances, telling others actually made the situation worse.

TFSV is a gendered problem that disproportionately impacts certain groups. It is important to keep in mind who is most at risk when developing TFSV resources and interventions.

We believe that with evidence-informed and co-ordinated action from the private and public sectors, young people can live in a digital world where they feel safe online and can easily access effective resources and support.

Charlotte Nau receives funding from the Social Sciences and Humanities Research Council of Canada.

Christopher Dietzel receives funding from the Social Sciences and Humanities Research Council of Canada.

Estefanía Reyes receives funding from the International Development Research Center (IDRC).

How to maintain good cognitive health at any age

2025-10-06T13:24:06Z

Is it an achievable goal to remain mentally sharp while aging, or is it a pipe dream?

It’s entirely possible if you cultivate habits throughout your life that are beneficial to brain function.

As a researcher in cognitive neuroscience and the neuropsychology of aging processes, I aim to shed light on the ways we can maintain good cognitive health while aging in light of recent scientific advances.

This article is part of our series The Grey Revolution. The Conversation Canada/La Conversation invites you to explore the impact of the aging boomer generation on Canadian society, including housing, working, culture, eating, travelling and health care … discover with us the upheavals underway, and those to come.

The importance of cognitive reserve

One of the most effective strategies research has identified is developing and maintaining good cognitive reserve.

Cognitive reserve refers to the brain’s ability to resist the effects of aging or neurodegenerative diseases without resulting in significant functional decline. This concept is now central to approaches for preventing cognitive decline.

In its report Dementia prevention, intervention, and care, updated in 2024, the Lancet highlighted the fact that 45 per cent of dementia cases could be prevented or delayed by addressing 14 modifiable risk factors.

These factors include physical inactivity, depression and social isolation.

But one of the earliest and most significant factors is having a low level of education.

Beyond education

Education has long been considered the main indicator of cognitive reserve. It reflects prolonged exposure to intellectually stimulating activities that promote the development of effective brain networks.

But this view is now considered incomplete. Cognitive reserve is not fixed in childhood or adulthood: it can be built, maintained and even amplified throughout life through different experiences including learning, rich social interactions and cognitively stimulating leisure activities.

Specific examples of these activities include playing a musical instrument or complex board games such as chess, or participating in volunteer activities that require planning and problem-solving skills.

Understanding cognitive reserve

Scientific research offers several complementary models for understanding the mechanisms of cognitive reserve.

Some focus on the structure of the brain itself, suggesting that characteristics such as the number of neurons influence the brain’s tolerance to damage. This is the brain reserve model, which is based on the idea that some people are born with a greater number of neurons, enabling them to cope better with aging.

Others argue that active lifestyles can slow down the effects of brain aging by strengthening biological resilience — for example, the brain’s ability to remain intact and functional as it ages, showing few visible signs of deterioration despite age. This is the brain maintenance model.

A third set of models emphasizes the functional flexibility of the brain, which allows it to mobilize its resources differently or recruit alternative neural networks to compensate for age-related losses. This is known as the cognitive reserve model.

These different models are part of a common conceptual framework that distinguishes between brain reserve, brain maintenance and cognitive reserve.

Each model is based on a specific idea, but they are complementary and supported by empirical data.

The cognitive reserve model remains the most widely studied, particularly because of its link to modifiable factors such as level of education and regular participation in cognitively stimulating activities.

Cognitive reserve is dynamic

This clarification helps to harmonize research and effectively guide prevention strategies. Above all, it reminds us that far from being a fixed entity, cognitive reserve evolves due to interactions with experience and learning, and can therefore be strengthened throughout life.

Recent work supports this dynamic view. A team of Québec researchers, of which I am a member, has shown that structured learning of memory strategies, including the method of loci (associating each piece of information with a familiar place) or mental visualization (transforming information into images to better retain it), can induce significant changes in brain activity.

A combination of increases and decreases in activation, including variations in the level of activity in different areas of the brain, was observed in different brain regions during the phases of learning and recalling information. This observation reflects the fact that the use of memory strategies allows for greater functional flexibility in the brain.

The results also showed that in more educated individuals, certain regions are activated in a more targeted manner during learning and recall, suggesting that their brains use more effective strategies.

Other research has also highlighted the role of education in brain structure and function. A study I conducted with colleagues highlighted an association between years of schooling, volume of grey matter and brain activation in the context of memory. Another study in which I participated showed greater flexibility of activation according to task complexity in more educated individuals.

All of this research confirms that cognitive reserve can be developed with experience and modulated by cognitive training at any age.

Stimulating your brain while having fun

In the same vein, the Engage study by the Canadian Consortium on Aging and Neurodegeneration aims to study the behavioural and neurophysiological effects of cognitively stimulating leisure activities in older adults.

This hybrid intervention combines formal cognitive training (memorization strategies, attention) with structured leisure activities such as learning music, a second language or video games.

It offers an ecological model — in other words, an approach that is close to real-life conditions, enjoyable and motivating and conducive to sustained engagement.

By demonstrating that these natural interventions produce effects comparable to those of traditional cognitive training programmes, which often consist of repetitive exercises done on a computer or on paper to work on functions such as memory or concentration, Engage could transform approaches to preventing age-related cognitive decline.

Learning another language

In my neuropsychology of aging laboratory (NeuroÂge) at the University of Québec at Trois-Rivières (UQTR), we are conducting a complementary project.

In collaboration with professors Paul John, from the Department of Modern Languages and Translation, and Simon Rigoulot, from the Department of Psychology, we are exploring the effects of learning English as a second language on cognition and brain activity in older adults.

Using a protocol that integrates classes, tutoring and cognitive and electroencephalography measurements, this project aims to document the cognitive and neural benefits of meaningful, motivating and accessible learning.

Preliminary results are promising and support the idea that intellectual engagement, even when started later in life, can generate measurable benefits.

Maintaining good cognitive health at any age requires a combination of accessible, motivating and stimulating interventions.

Cognitive reserve, far from being fixed, is built up throughout life. Advances in research now offer us concrete tools for healthy aging, particularly when it comes to cognitive health.

Benjamin Boller received funding from the Fonds de Recherche du Québec.

Politically aggressive social media users are creating most of the anti-immigrant content

2025-10-05T13:26:01Z

Most of us, whether we admit it or not, engage in a great deal of passive scrolling through social media daily.

And while the platforms have proliferated for years, experts are only now beginning to demonstrate their full impact on our attention, mental health, spending habits and politics.

Despite the benefits, social media is also creating new problems. A pressing concern is the dissemination of misinformation by political extremists, a trend amplified by the unprecedented reach of platforms like Facebook and X (formerly Twitter). When it comes to issues like immigration, many activists, experts and pundits point to social media as a vehicle for the spread of prejudice, conspiracy theories and false claims targeting immigrant and minority populations.

Even before launching his 2016 presidential bid, for example, Donald Trump used Twitter to share messages attacking immigrants and ethnic minorities with millions of people, giving him the power to dominate news cycles and shape public policy.

Does social media make people more xenophobic?

Polarizing platforms

For decades, scholars studying how people consume information about immigration have argued that print and TV news stories often portray the economic and social impact of immigration negatively.

Studies on major American newspapers and news stations show that traditional media coverage has encouraged prejudice toward Latin American immigrants and Muslims.

Does social media follow this trend? Social scientists are beginning to disagree.

Some scholars point to racist and anti-immigration messages on social media as evidence that platforms like Facebook, X and Reddit encourage users to speak freely without the constraints of social norms to a broad and diverse audience.

Other studies argue that social media creates uniquely polarizing environments where users organize themselves into political tribes that fight one another using aggressive dialogue. Even in Canada — a country often touted as pro-immigration — social media has allowed users to attack immigrants and minorities.

Users’ attitudes, however, may matter more than the specific platform.

Politically aggressive users

Recent studies from the United States and Western Europe show that social media attracts politically aggressive users who often do most of the talking in heated online conversations.

Based on my recent research on Canadian X users, I found similar results. I analyzed roughly 13,000 English-language posts discussing immigration and Canada’s housing crisis in 2023. Unsurprisingly, I discovered that many users blamed immigrants for a lack of affordable housing, including influencers with tens of thousands of followers.

In August 2023, discussions about housing on X peaked, with 3,638 posts mentioning both immigration and housing. This significant increase in online conversation coincided with federal government’s public comments linking international students to the housing crisis. The data supports the idea that Canadians were actively discussing the housing crisis in relation to immigration during this time.

Does this mean that Canadian X users are now seething with hatred for immigrants? While some are, a closer look reveals the partisan nature of these posts.

When I examined users’ identities and networks, it became clear that their anti-immigration messages were often a means of criticizing Justin Trudeau and his Liberal government. In other words, right-wing users (with large and small followings) were chiefly responsible for creating and sharing these posts, including People’s Party of Canada leader Maxime Bernier.

For instance, Fringe Albertan (about 2,500 followers in August 2023) posted in response to a post by Rebel News:

“@RebelNewsOnline Its a lie! Typical Liberal. Hes lying bc Canada is a UN member, and as a member, has signed onto an immigration pact to flood Canada with migrants, destroying our economy, social network, housing, and culture. #EndUNMembership @UCPCaucus @CPC_HQ @Buffalo_AB @BuffaloPartySK”_

Similarly, lloyd (about 50 followers at the time) posted in response to a post by CTV News:

“@CTVNews Thanks CTV News it’s no wonder why they are leaving as Canada is so poorly governed ! Housing shortage when Immigration brings millions of Migrants and never checked to see how many homes they had and shortage worst ever for Canada! Worst blunder in Canadian History! HELP.”

Right-wing social media users significantly contributed to public discourse blaming immigrants for Canada’s problems.

Some might argue polarizing content is simply a reflection of free speech.

This is true to some degree, but recent studies suggest online polarization can also threaten free societies. Algorithms designed to focus users’ attention on threats and conflict can reliably make users engage with content; this is what makes social media platforms potentially dangerous. Fortunately, users are far from powerless.

Reducing online polarization

While figures like Trump show that social media can be used to spread prejudice to mass audiences, it also matters that users often self-select into networks they like.

New studies make clear that users’ socio-political context, partisanship and behaviour seem to matter as much as the platform itself.

It turns out both platforms and users are responsible for online polarization.

What can we do about social media platforms?

Ultimately, we need socially responsible online platforms that focus less on producing outrage and division to attract users. This means including researchers, governments and civil society in designing social media interfaces and algorithms to establish reasonable community standards for sharing information and regulating users’ behaviour.

But we cannot wait for politicians to solve this problem. Even if we get platforms that focus less on outrage, trolls will still exist.

Social media’s rapid pace and the lack of consensus over online behaviour create ethical dilemmas for users everywhere. For example, many people passively scroll and react to content they skimmed, but if conflict arises later in the thread, many users are unsure how to respond or whether they should respond at all.

To see less polarizing social media content, we need to both consciously choose what platforms we wish to join (and why), and we need to cultivate better ways to handle online conflict.

Nicholas A. R. Fraser does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

YouTube shapes young people’s political education, but the site simplifies complex issues

2025-10-05T13:25:16Z

There is a widely held misconception that young people are politically disengaged. This is based on narrow measures like voter turnout. But this overlooks the fact that many young people are deeply curious, especially when politics is understood more broadly: as a way to make sense of society, power and everyday life.

In my research with youth in the Greater Toronto Area, I explored how their views on inequality, identity and government form over time.

My findings build on my earlier research, conducted with second-generation Chinese and South Asian Canadians, where I found that many of them turned to conservative ideas to access feelings of dignity and belonging. For them, embracing meritocracy wasn’t about denying racism — it was a way to prove they’d succeeded by Canada’s rules.

In this new study, I wanted to understand what shapes that gap — what makes some students more likely to see power as structural, and others more likely to see it as personal or cultural.

I found that young people now form political beliefs through two competing knowledge systems: a hollowed-out university, and YouTube’s attention economy. In the university classroom, students learn to connect experience to systems like racism or class inequality. On YouTube, other students encounter simplified stories or common-sense clichés.

The result is a generation pulled between critique and clarity, where YouTube offers answers that feel true.

Changes to postsecondary education

Post-secondary institutions in Canada have historically played a central role in public life. They offered young people a place to explore political ideas, learn history and develop critical thinking skills. That mission has since eroded.

In Ontario, former premier Mike Harris’s so-called “Common Sense Revolution” marked a turning point in government approaches to education. Post-secondary education was rebranded as an individual investment rather than a public good. The cost of tuition increased, public funding stagnated and student debt rose.

As a result, academic paths became stratified. Lower-income students pursued vocational degrees, while their wealthier peers could afford less lucrative paths, like the social sciences and humanities.

The ability to encounter transformative ideas narrowed along class lines.

Market priorities

At the same time, disciplines like sociology and history began to lose institutional standing as universities became increasingly reliant on tuition fees, corporate partnerships and research tied to economic outcomes. Funding shifted toward programs seen to deliver market returns — like business and technology — while fields focused on critique or public interest were sidelined.

This reorientation entrenched the idea that higher education exists to serve the market. So it was no surprise when Ontario announced $750 million in new post-secondary funding; none for the social sciences and far below the $2.5 billion recommended by a government-appointed group tasked with reviewing the financial sustainability of Ontario’s post-secondary system.

Universities are now judged by job outcomes for graduates, with less support for courses that analyze, critique or challenge inequality or power.

YouTube steps in

As universities retreat, platforms like YouTube have increasingly stepped in as a political educator. This is accelerating a shift that may have happened anyway, but has now taken on a new urgency in this hollowed-out educational landscape.

In 2015, YouTube’s algorithm shifted to maximize watch time, pushing content independent of its quality.

I found that for students in technical or vocational programs — where inequality is rarely addressed — YouTube often becomes their main source of political learning.

Conservative influencers offer simplified narratives: inequality reflects natural differences, tradition ensures order, progressivism is elitist.

These messages land because progressive ideas remain concentrated in universities, out of reach for many working-class youth. This dynamic has also expanded across platforms like TikTok and Instagram, where short-form content delivers similar emotionally charged explanations.

When critical education is confined to elite spaces, structural thinking becomes a privilege and not a public tool. This matters because it shapes who feels entitled to analyze power and imagine alternatives — and who is left to make sense of inequality through personal experience and YouTube algorithms.

Some young people are pushing back: BreadTube creators, civic tech projects and public sociology podcasts translate progressive ideas for digital audiences.

But these efforts remain small compared to the reach and resources of right-wing media. Without broader infrastructure — from education funding to algorithmic transparency — even the most compelling content struggles to shift how people understand the world.

The decline of progressive institutions

Universities were never the only sites of political education. In earlier generations, unions, political parties and community groups shaped public consciousness.

They established adult education programs, published newspapers and linked political ideas to everyday life. Feminist and anti-racist traditions added their own spaces, from women’s consciousness-raising circles to Black political study groups.

Civic initiatives like Company of Young Canadians, supported youth in under-served communities with political engagement and collective action. These institutions helped working people identify shared interests and organize for change.

That world has largely disappeared, especially with the decline of unions in Canada, driven by decades of neoliberal restructuring that weakened collective bargaining and eroded political education.

In Canada, the New Democratic Party has increasingly prioritized electoral success over grassroots organizing. This isn’t unique to the NDP. Across the North America, left-leaning organizations often function as symbolic communities, struggling to build collective power.

Their abstract language feels out of step with people navigating material problems like rent hikes and job precarity.

In contrast, the political right speaks plainly. And, its messages may be simple, but they are easy to find.

Cultivating critical thought

When universities retreat and progressive organizations lose influence, new forces shape how people come to understand the world.

My research found that the way Canadian youth explained inequality differed depending on their access to education. Students with post-secondary social science education connected personal experience to systemic inequality. Those outside these spaces — especially those relying on YouTube — were more likely to see inequality as natural, rooted in individual effort or cultural values.

This divergence reflects a deeper shift: the pipeline for developing structural literacy has broken down. Where critical thinking was once nurtured through unions, political parties and public education systems, those institutions have thinned out.

With unions weakened and parties consumed with electoral success, the university remains one of the few institutions still cultivating critical thought — and conservative leaders know it.

Ahead of the 2025 election, the Conservative Party pledged to end the “imposition of woke ideology” in university research funding and steer university hiring “away from ideology.”

For from neutral, these efforts turn universities into places where challenging ideas are no longer welcome. In their place, young people are left to navigate politics through platforms shaped by algorithms, where nuance is rare.

A different future

If we want a different future, where more people feel equipped to understand and change the world, we need institutions that foster imagination, inclusion and collective purpose. That means rebuilding unions, community-based groups and civic networks.

It also means rethinking what political parties and universities are for.

Political parties must organize, not just campaign. Universities must educate for democratic participation, not just employability. These institutions must do the slow, relational work of building solidarity: helping people understand the systems they live in and feel part of something larger than themselves.

Without that kind of infrastructure, progressive ideas stay abstract: visible to some, but disconnected from the everyday lives of most.

Emine Fidan Elcioglu does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

The American TikTok deal doesn’t address the platform’s potential for manipulation, only who profits

2025-10-02T15:26:08Z

On Sept. 25, the Donald Trump administration in the United States again extended the TikTok ban-or-divest law, possibly for the last time. The latest extension to the law, which was passed in 2024 by the Joe Biden administration, includes a deal to transfer TikTok to American owners as a condition required to avoid a ban.

This raises the question on the validity of the warnings about the app as a tool of Chinese influence and whether American ownership will help.

Canada should be watching closely, because anxieties about foreign manipulation and social media exist north of the border, too. These range from bans on TikTok and concerns about Beijing-linked surveillance to efforts like Bill C-18 aimed at safeguarding domestic news sources.

What happens in the Canadian information environment has always been shaped by the U.S., a dependence that is even more precarious now that American politics has turned hostile to Canada.

ABC News covers the executive order that brought into effect U.S. ownership of TikTok.

TikTok concerns

TikTok is not the only digital media platform susceptible to worries about hostile influence. All major platforms introduce the same vulnerabilities. If the policy objective is to enhance the security of democracy, then a focus on TikTok is too narrow and divestment as a solution accomplishes little (especially because it appears China will retain control of the algorithm).

Worries about TikTok come down to two big fears. The first is that it functions as a spying machine, feeding data to the Chinese government. The spying concern isn’t just about espionage, learning about sensitive infrastructure and activities, but also personal — the software itself might be unsafe and can be used to track individuals.

As a result, many countries have banned the app on government devices, and securing data along national borders may well address this.

The second fear, more vivid in the public and political imagination, is that TikTok functions as an influence machine. Its algorithm can be tweaked to push propaganda, sway opinion, censor views or even meddle in elections.

Such worries reached a fever pitch in America in 2023, when Osama bin Laden’s “Letter to America” suddenly went viral on TikTok. Lawmakers seized on this as evidence that TikTok could amplify extremist content, reinforcing fears that the platform can be weaponized.

These worries aren’t merely speculative. Investigations have shown that topics sensitive to China, such as Tiananmen Square and Tibet, are harder to find or conspicuously absent on TikTok compared to other platforms.

Social media is also used as a tool for influence by hostile groups, corporations and governments, and concerns about ownership are often a proxy for deeper anxieties about the platforms themselves.

As users, we know little about how our feeds work, what’s shaping them, what they might look if they were built differently and how they are affecting us.

There is a rational basis to be mistrustful, and this cuts both ways. It’s not just the fear that we could be manipulated without realizing it; it’s also the temptation to see our opponents as manipulated, too, as if every disagreement might be product of someone rigging the system.

Users know little about how TikTok feeds work, what’s shaping them or what they might look if they were built differently. (Solen Feyissa/Unsplash), CC BY

Manipulated anxieties

Fear of TikTok as an influence machine continues to play a substantial role in politics, as “Washington has said that TikTok’s ownership by ByteDance makes it beholden to the Chinese government.”

U.S. Vice President JD Vance remarked that the executive order would “ensure that the algorithm is not being used as a propaganda tool by a foreign government… the American businesspeople … will make the determination about what’s actually happening with TikTok.”

Meanwhile, Trump ostensibly joked that he’d make TikTok “100 per cent MAGA” before adding “everyone’s going to be treated fairly.” And Israeli Prime Minister Benjamin Netanyahu told an audience of content creators that “weapons change over time… the most important one is social media,” stressing the importance of divestment of TikTok to U.S. owners.

One implication of these comments is that divestment doesn’t change the threat of manipulation — it just changes who’s doing the manipulating. Divestment is framed as resisting foreign propaganda, but at the same time domestic manipulation is legitimized as politics as usual.

Collective dependence

This is a squandered opportunity for the U.S. By treating TikTok as a weapon to be seized, leaders have passed up the chance to model a more enduring form of soft power: building open, transparent, trustworthy information systems that others would want to emulate. Instead, what is gained is a temporary and possibly illusory sharp power advantage, at the expense of an enduring source of legitimacy.

The bigger problem is that the normalization of social media as a weapon is, to borrow a fear familiar to Trump, riggable. We know that social media can be manipulated, and yet we rely on it more and more as a source of news. And even if we ourselves don’t, we are influenced indirectly by those who do.

This collective dependence makes the platforms more powerful and their vulnerabilities more dangerous.

Social media platforms have become a primary source of information. (Shawn/Unsplash), CC BY

Protecting the public sphere

Canada has already had its own TikTok moment: the Online News Act (C-18), which required platforms to pay news outlets for sharing their content. This was intended to strengthen Canadian journalism, but in response, Meta banned news on its platforms (Facebook, Instagram) in Canada in August 2023, leading to an 85 per cent drop in engagement. Instead of strengthening Canadian journalism, Bill C-18 risks making it more fragile.

If we’re serious about protecting the public sphere from manipulation, what matters is the outsized power the platforms have, and the extent to which that power can be bought, sold or stolen. This power includes the surveillance power to know what we will like, the algorithmic power to curate our information diet and control of platform incentives, rules and features that affect who gains influence.

Bargaining with this power, as Canada tried with Bill C-18 — and as the U.S. is now doing with China and TikTok — only concedes to it. If we want to protect democratic information systems, we need to focus on reducing the vulnerabilities in our relationship with media platforms and support domestic journalism that can compete for influence.

The biggest challenge is to make platforms less riggable, and thus less weaponizable, if only for the reason that motivated the TikTok ban: we don’t want our adversaries, foreign or domestic, to have power over us.

Andrew Buzzell does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Why we should be skeptical of the hasty global push to test 15-year-olds’ AI literacy in 2029

2025-09-30T13:31:13Z

If 2022 was the year OpenAI knocked our world off course with the launch of ChatGPT, 2025 will be remembered for the frenzied embrace of AI as the solution to everything. And, yes, this includes teaching and schoolwork.

In today’s breakneck AI innovation race, the Organization for Economic Co-operation and Development (OECD), along with the European Commission, have called for the development of unified AI literacy strategies in kindergarten to Grade 12 education.

They have done this through an AI Literacy Framework developed with Code.org, and a range of experts in computational thinking, neuroscience, AI, educational technology and innovation — and with “valuable insights” from the “TeachAI community.”

The “TeachAI community” refers to a larger umbrella project providing web resources targeting teachers, education leaders and “solution providers”. Its advisory committee includes companies like Meta, OpenAI, Amazon and Microsoft and other for-profit ed tech providers, international organizations and government educational agencies and not-for-profit groups.

The rush to establish global standards for AI literacy has been further energized by a recent OECD program announcement.

The Programme for International Student Assessment (PISA) — which tests 15-year-old students of member nations in literacy, numeracy and science every three years — is introducing a media and AI literacy assessment in 2029. This is related to what it calls an “innovation domain” of learning.

There have been consultations about the AI literacy framework, but it’s misguided to think that educators and the general public at large would be able to comment on this in an informed way before AI has been widely accessible to the public.

The OECD’s hasty push for PISA 2029 threatens to obscure essential questions about the political economy that is enabling the marketing and popularization of AI, including relationships between business markets and states.

Marketing, popularizing AI

Essential questions include: Who stands to benefit most and profit from proliferating AI in education? And what are the implications for young people when national governments and international organizations appear to be actively promoting the interests of private tech companies?

We agree with a growing community of researchers that regard calls for AI literacy as being based on ill-defined and preliminary concepts: for example, the draft framework speaks about four areas of AI literacy competency that involve: engaging with AI, creating with AI, managing AI and designing AI.

As we try to grasp the meaning of terms such as “AI skills” and “AI knowledge,” the educational landscape becomes both vague and confounding. Educators are all too familiar with the legacy, often related to commercialization, of attaching various modifiers to notions of literacy — digital literacy, financial literacy, the list goes on.

‘The future’

By framing AI as a distinct, readily measurable capability, the OECD has signalled that it can impose its own understanding onto AI, leaving school communities globally with the task of simply accepting and implementing this presumed all-embracing vision of the future amid profound and alarming existential and practical questions.

Efforts to frame AI literacy as a vehicle to prepare young people for “the future” are a recurring theme of influential global policy bodies like the OECD.

Elsewhere, research has shown how these policy shifts over the past three decades follow a familiar pattern — the OECD functions as an influential policy entity that establishes its own definitions of student progress through standards and benchmarks for assessing the quality of education programs around the globe. In doing so, it imposes a single understanding on what are diverse systems with distinct cultures.

As digital education expert Ben Williamson points out, this burst of “infrastructuring AI literacy” not only involves “building, maintaining and enacting a testing and measurement system” but will also “make AI literacy into a central concern and objective of schooling systems.”

In doing so, it will sideline other important subjects, gear up schools and learners to become uncritical users of AI and turn schools into a testing ground for AI developments.

Lack of discussion around teachers

We also have other concerns.

In our preliminary research, yet to be published, we analyzed the AI Literacy Framework document and found a significant lack of discussion regarding the role of teachers. The document directly mentions teachers only 10 times and schools nine times. By comparison, AI is mentioned 442 times, while learners and students are referenced approximately 126 times.

This suggests to us that teachers and formal schooling seem to have been removed from any major role in these frameworks. When they are mentioned, they appear a more of a prop to AI and not a critical mediator.

Educators and national education systems are facing a one-size-fits-all solution to a wider societal issue that attempts to defuse, depoliticize and naturalize what ought to be urgent, engaged conversations by teachers and the education profession about AI, education, learning, sustainability and the future.

Current classroom realities

As political theorist Langdon Winner reminded us more than 40 years ago, technologies have politics that rotate around both problems and opportunities. These politics ignore some realities and amplify others.

Well-intended promoters of AI literacy in schools in Canada call for professional development and resources to support the adoption of AI. Yet these aspirations and hopes for positive change need to be contextualized by the current realities Canadian teachers face:

63 per cent of educators report their ministries of education are “not supportive at all;”
Nearly 80 per cent of educators report struggling to cope;
95 per cent of educators are concerned that staff shortages are negatively impacting students.

Proceed with slowly with care

Ours is not a call for educators to be luddites and reject technology. Rather, it’s a call to the profession and the public to collectively question the rush to AI and the current framings of AI literacy as an inevitable policy trajectory and preferred future for education.

Both the limited time frame of the next few months to respond to the AI Literacy Framework — following its May 2025 release — and the pre-emptive decision by the OECD to proceed with its PISA assessment in 2029 signals a race to a finish line.

As with the recent return to school and the annual reminders about the need for caution in school speed zones, we need to avoid distractions — and proceed slowly, with care.

Michele Martini received funding from the European Research Council (Grant agreement No. 837727)

J-C Couture and Susan Lee Robertson do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.

Our AI model can help improve indoor ventilation during wildfire season

2025-09-30T13:22:12Z

A recent report from the University of Chicago’s Air Quality Life Index found that wildfires are worsening air quality in Canada. The report found that in 2023, wildfires caused concentrations of particulate matter to rise to levels not seen since the index started taking records in 1998.

This summer, Canada experienced one of the worst wildfire seasons on record. Fires caused thousands to evacuate their homes and smoke periodically blanketed cities, causing outdoor air quality to deteriorate.

When we smell or see smoke, the first thing many of us might think to do is close our windows. However, wildfire smoke contains small fine particulate matter (PM2.5) that can pass through small openings or gaps.

In 2023, wildfires in Canada caused more greenhouse gas emissions than all other sources combined. That means designing safer indoor spaces is a public health imperative. But how can we develop indoor spaces that are well-ventilated and safe from the harmful effects of smoke?

Enhancing indoor air quality

Answering this question would traditionally require going through a real-world process of trial and error in various spaces. Such a process is time-consuming and not always feasible. However, we recently developed a framework integrating modelling and simulation with deep learning techniques to help answer this question.

We know that enhancing indoor air quality, whether through improved ventilation, an optimal occupancy-to-area ratio or other room setting adjustments, can improve health and reduce the spread of infections.

The next step for researchers and designers is to determine the best indoor design features to reduce carbon dioxide concentration. Such features include rooms dimensions, the location of ventilation ports, ventilation levels, where windows are, maximum number of occupants, seating arrangements and so on.

How our model works

Our framework tackles two pertinent problems: the lack of verified, accurate information and the inefficiency of producing and studying simulation results for many combinations of settings.

We use an advanced mathematical model and associated software tools that allow us to simulate varied enclosed spaces with different settings, and to collect simulation results.

The simulated data is then further used to form a data set to train an AI algorithm — in this case, using a deep neural network. Designers can use the trained network to predict unknown settings of the closed space when other settings are altered.

The framework allows designers to simulate how changes in room layout, such as the number vents and where they are placed, or the density of occupants, could impact well-being. For example, the framework can estimate how many people might get sick in a given space, helping architects and planners adjust configurations to minimize infection risk before construction begins.

We used several case studies from university laboratory settings to validate the framework. In one case study, our research team could create 600 simulation scenarios of different laboratory designs. The simulation results produced a rich dataset that would be nearly impossible to replicate in real life due to cost and logistical constraints.

The resulting dataset is used to train a machine learning algorithm to predict where and how many people might be exposed to high levels of carbon dioxide. With that information in hand, it’s easier to make smart decisions about where to place ventilation ports or how many people should safely occupy a room under specific conditions.

Future studies needed

Across Canada, researchers are leveraging machine learning to study indoor air quality in homes, schools and offices. Our findings suggest that this approach is well-suited for studying how carbon dioxide spreads in indoor environments.

However, broader study is still needed. To date, case studies have focused exclusively on a university environment. Yet our framework is designed to be scaleable and adaptable to wide range of indoor spaces. Future research should expand to schools, gymnasiums and residential buildings to strengthen the trust in the framework and refine its predictive power.

As climate change intensifies wildfire seasons, Canadians will spend more time indoors avoiding smoke. The good news is that we have the tools, data and the scientific insight to make indoor spaces healthier and safer for everyone.

We may not have the means to control the air outside, but we can design our spaces to control the quality of the air inside.

The authors do not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.

Generative AI might end up being worthless — and that could be a good thing

2025-09-28T17:03:48Z

In the rush to cash in on the generative artificial intelligence gold rush, one possible outcome of AI’s future rarely gets discussed: what if the technology never works well enough to replace your co-workers, companies fail to use AI well or most AI startups simply fail?

Current estimates suggest big AI firms face a US$800 billion dollar revenue shortfall.

So far, genAI’s productivity gains are minimal and mostly for programmers and copywriters. GenAI does some neat, helpful things, but it’s not yet the engine of a new economy.

It’s not a bad future, but it’s different from the one currently driving news headlines. And it’s a future that doesn’t fit the narrative AI firms want to tell. Hype fuels new rounds of investment promising massive future profits.

Maybe genAI will turn out to be worthless, and maybe that’s fine.

Indispensable or indefensible?

Free genAI services, and cheap subscription services like ChatGPT and Gemini, cost a lot of money to run. Right now, however, there are growing questions about just how AI firms are going to make any money.

OpenAI CEO Sam Altman has been candid about how much money his firm spends, once quipping that every time users say “please” or “thank you” to ChatGPT, it costs the firms millions. Exactly how much OpenAI loses per chat is anyone’s guess, but Altman has also said even paid pro accounts lose money because of the high computing costs that come with each query.

Like many startups, genAI firms have followed the classic playbook: burn through money to attract and lock-in users with a killer product they can’t afford to miss out on. But most tech giants have not succeeded by creating high-cost products, but rather by making low-cost products users can’t quit, largely funded by advertising.

When companies try to find new value, the result is what journalist and author Cory Doctorow coined “enshittification,” or the gradual decline of platforms over time. In this case, enshittification means the number of ads increase to make up the loss of offering the free service.

OpenAI is considering bringing ads to ChatGPT, though the company says it is being “very thoughtful and tasteful” about how this is done.

It’s too soon to tell whether this playbook will work for genAI. There is a possibility that advertising might not generate enough revenue to justify the massive spending needed to power it. That is because genAI is becoming something of a liability.

The hidden costs of AI models

Another looming problem for genAI is copyright. Most AI firms are either being sued for using content without permission or entering costly contracts to licences content.

GenAI has “learned” in a lot of dubious ways, including reading copyrighted books and scraping nearly anything said online. One model can recall “from memory” 42 per cent of the first Harry Potter novel.

Firms face a big financial headache of lobbying to exempt themselves from copyright woes and paying off publishers and creators to protect their models, which might end up a liability no matter what.

American AI startup Anthrophic tried to pay authors around US$3,000 dollars per book to train its models, adding up to proposed settlement that added up to US$1.5 billion dollars. But it was quickly thrown out by the courts for being too simple. Anthrophic’s current valuation of US$183 billion might get eaten up pretty quick in lawsuits.

The end result of all this is that AI is just too expensive to be owned, and is becoming something like a toxic asset: something that is useful but not valuable in and of itself.

Cheap or free genAI

Meta, perhaps strategically, has released its genAI model, Llama, as open source. Whether this was meant to upset its competitors or signal a different ethical stance, it means anyone with a decent computer can run their own local version of Llama for free.

Open AI models are another corporate strategy to lock in market share, with curious side effects. They are not as advanced as Gemini or ChatGPT, but they are good enough, and they are free (or at least cheaper than commercial models).

Open models upset the high valuations being placed on AI firms. Chinese firm DeepSeek momentarily tanked AI stocks when it released an open model that performed as well as the commercial models. DeepSeek’s motives are murky, but it’s success contributes to growing doubts about whether genAI is as valuable as assumed.

Open models — these by-products of industrial competition — are ubiquitous and getting easier to access. With enough success, commercial AI firms might be hard pressed to sell their services against free alternatives.

Investors could also become more skeptical of commercial AI, which could potentially dry up the taps of seed money. Even if open access models also end up being sued into oblivion, it will be much harder to remove them from the internet.

Can AI ever be owned?

The idea of genAI being worthless might recognize knowledge is intangibly valuable. The best genAI models are trained off the world’s knowledge — so much information that the true price may be impossible to calculate.

Ironically, these efforts by AI firms to capture and commercialize the world’s knowledge might be the thing damning their products; a resource so valuable a price cannot be attached. These systems may be so indebted to collective intellectual labour such that their outputs cannot truly be owned.

If genAI can’t generate sustainable profits, the consequences will likely be mixed. Creators pursuing deals with AI firms may be out of luck; there will be no big cheques from OpenAI, Anthropic or Google if their models are liabilities.

Progress on genAI could stall, too, leaving consumers with “good enough” tools that are free to use. In that scenario, AI firms may become less important, the technology a little less powerful — and that might be perfectly OK. Users would still benefit from accessible, functional tools while being spared from another round of overhyped pitches doomed to fail.

The threat of AI being worth less than anticipated might be the best defence against the growing power of big tech today. If the business case for generative AI proves unsustainable, what better place for such an empire to crumble than on the balance sheets?

Fenwick McKelvey receives funding from the Social Sciences and Humanities Research Council of Canada and the Fonds de Recherche du Québec.

The warning signs are clear: We’re heading toward a digital crisis

2025-09-25T12:47:54Z

People’s lives are more enmeshed with digital systems than ever before, increasing users’ vulnerability and insecurity. From data leaks like the 2017 Equifax data breach to the more recent cyberattack on British retailer Marks & Spencer, business operations and data on the internet continue to be vulnerable.

There are good reasons to believe that little will be done about these risks until a massive society-wide crisis emerges.

My research suggests that there are significant failures in our current approaches to risk and innovation. Digital technologies remake social life through new technologies, communication platforms and forms of artificial intelligence. All of which, while very powerful, are also highly risky in terms of malfunctioning and vulnerability to being manipulated.

Yet, governments are generally unable to distinguish between what are actually valuable contributions to society and what are intensely socially damaging.

CBC’s The National looks at data breaches.

A massive social experiment

The digital economy includes “those businesses that increasingly rely upon information technology, data and the internet for their business models.” The companies dominating the digital economy continue to undertake a massive social experiment where they keep the lion’s share of the benefits while shunting the risks onto society as a whole.

This could lead to a systemic digital crisis, ranging from a widespread breakdown of basic infrastructure, such as electricity or telecommunications due to a cyberattack, to an attack that modifies existing infrastructure to make it dangerous.

There are significant similarities between the current trajectory of the digital economy and the 2008 financial crisis. In particular, what we are increasingly seeing in the digital world, which we saw in the pre-crisis financial world, is what American sociologist Charles Perrow called “tight coupling.”

Perrow argues that when systems exhibit high levels of interconnection without sufficient redundancy to compensate for failures, it can lead to catastrophic consequences.

Likewise, high levels of complexity are generally considered to make highly interconnected systems riskier. Unanticipated risks and connections can lead to failures cascading across the system.

Increasing interdependence

Our existing digital economy shares many of these characteristics. The digital economy is characterized by a business model that focuses on businesses getting as large as possible as quickly as possible.

The lead-up to the 2008 financial crisis and the current digital economy share both the amplification of interdependency alongside the reduction of redundancy. In the case of finance, this proceeded through massive borrowing to leverage earnings, leaving a smaller ratio of money left to cover any possible losses.

In the digital economy, this need to continually collect data increases interdependencies among datasets, platforms, corporations and networks. This increased interdependency is fundamental to the core business model of the digital economy.

The undermining of redundancy in the digital sphere is manifested in the “move-fast-and-break-things” ethos in which digital companies eliminate or acquire competitors as quickly as possible while eliminating analog alternatives to their own digital networks.

Last, these digital behemoths and their rapid growth increase the complexity of the digital economy and the monopolistic networks that dominate it.

BBC News covers last summer’s flight cancellations.

Obvious warning signs

There is a key difference between the 2008 financial crisis and the contemporary digital economy. Unlike in the lead-up to the crisis, where a partially finance-driven prosperity quieted any obvious warning signs, the warning signs in the digital economy are front and centre for everyone to see.

The 2017 WannaCry and NotPetya malware attacks each caused billions of dollars in damages. More recently, the CrowdStrike failure in 2024 cancelled thousands of flights, and even took television stations off the air. Constant hacks, ransomware attacks and data leakages are warning signs that this is a deeply fragile system.

AI has taken many of these vulnerabilities into overdrive, while adding new risks, such as AI hallucinations and the exponential growth in misinformation. The speed and scale of AI are expected to intensify existing risks to confidentiality, system integrity and availability.

This is potentially the most significant, though unfortunate element in this story. There is massive system risk, yet they are not addressed directly, and the processes heightening these risks continue to accelerate.

This suggests a deeper problem in our politics. While we do have some ability to regulate after the damage is done, we struggle to prevent the next crisis.

Dean Curran received funding from the Social Sciences and Humanities Research Council of Canada.

Gen Z protests brought about change in Nepal via the powers — and perils — of social media

2025-09-25T12:46:34Z

Youth protesters in Nepal are in the global spotlight for their angry response to the government’s sweeping social media ban in an apparent attempt to silence their dissent. The government’s actions ignited mass protests — led largely by Gen Z, a cohort made up of young people born between 1997 and 2012.

The Gen Z movement represents a turning point for Nepal politics. The protesters had three key demands: end corruption, end nepotism and reform the country’s political systems.

Their uprising led to the resignation of Prime Minister K.P. Sharma Oli and several government ministers. Sushila Karki was then appointed interim prime minister, and the protests have since died down.

Why is the Gen Z protest unique?

Nepal’s Gen Z movement is different from other movements in Nepal.

First, it is led by young people. Second, social media is their main means of communicating their dissent and their agenda.

These protesters are angry that working-class young people are struggling to meet basic everyday needs (food, shelter, jobs, health care, etc.) and facing rising inequality, discrimination and poverty.

That’s in contrast to the children and grandchildren of Nepal’s high-profile elite politicians, accused by the protesters of living in the lap of luxury. Gen Z protesters have demanded information about the source of income of Nepal’s ultra-rich politicians and their families, and called for a thorough investigation.

A segment on how Nepal’s Gen Z protesters targeted #nepokids. (Sky News)

Why are Gen Zs so frustrated?

For a long time, Nepal, with a population of 29.5 million, has been trapped in a poverty cycle. It is ranked 143rd globally in the Human Development Index (2024).

The unemployment rate in 2024 for youth aged 15-24 was 20.82 per cent, and it’s growing. Reports also suggest that more than 1,500 adults leave the country every day in search of work.

In 2021, the Nepal census found that 7.1 per cent of the population was working outside the country and has a median age of 28.

In 2023, Nepali workers sent remittances of US$11 billion back home. In fact, estimates suggest that almost 25 per cent of Nepal’s GDP is from remittances.

There is also growing concern about Nepali worker deaths as people take dangerous jobs; more than 700 workers died from 2018 to 2019 Gen Z frustrations are linked to how their parents leave the country in search of work and do the most high-risk and lowest-paid jobs abroad, which they believe is in stark contrast to the lives of #NepoBabies and #NepoKids.

Gen Z’s digital tactics

Some Gen Z social media users tracked the accounts (on Instagram, TikTok and Facebook) of the children and grandchildren of ultra-rich politicians and shared or reposted images and videos of their luxurious lifestyles.

That included photos taken on high-end vacations in Europe, shopping for designer brands like Louis Vuitton, Prada, Gucci and Cartier, as well as their stays in family properties worth billions.

Social media engagement surged on posts with these images and with hashtags that included #Nepobaby, #NepoKids, #PoliticiansNepoBabyNepal and #Corruption.

Some Gen Zs also made short videos on TikTok and Facebook highlighting corruption, inequality, poverty and nepotism; those videos also went viral.

All of these issues resonated with many Nepali Gen Zs, spurring them to join the protest movement.

Social media ban

Before Sept. 8, Gen Z’s protests were peaceful and mostly took place online. But when the government instituted a ban on social media, Gen Z erupted, with many claiming that the decision was aimed at silencing their voices.

Gen Z is the social media generation, and the ban was regarded as a violation of their rights. They soon took their demands to the streets from the screen, calling for the resignation of the prime minister.

The protest turned into a battlefield as police killed 19 school-aged students on the same day; hundreds were also injured. As of now, the Gen Z protester death toll is 72.

Aftermath

The prime minister resigned on Sept. 9, but the situation further worsened. Protesters burned down key government buildings, including parliament and court buildings, private businesses, banks and the homes of politicians and business people across the country.

After a series of talks between the chief of the Nepal army, Ashok Raj Sigdel, Nepali President Ram Chadra Poudel and Gen Z leader Sudan Gurung, an interim six-month government was formed. Karki was appointed the first female prime minister of the country.

The interim cabinet’s priorities include the upcoming election in March 2026, tackling corruption, investigating the killings of Gen Z protesters as well as the destruction of public and private property.

The power and perils of social media

Before the Nepal protests, dissenting youth in countries that include Bangladesh, Sri Lanka and Myanmar have used social media to air their grievances.

A study has shown how social media plays a role in empowering youth, amplifying marginal voices and building transnational solidarity. Examples include some of the most popular global social movements like #BlackLivesMatter, #MeToo, #MahsaAmini.

But its role in protest movements can also be problematic.

Amid the Gen Z protests in Nepal, reports of disinformation and misinformation are spreading. A video claiming 35 human skeletons were found in a store was posted on Sept. 13 by a Facebook user with 63,000 followers, fuelling panic among the protesters. The claim was determined to be false.

Gen Z protesters in Nepal and beyond are clearly having some success in bringing about social and political change. But with the growth of artificial intelligence, creating fake content is no longer difficult, and false information can proliferate quickly amid this generation.

Luna KC does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

How researchers are making precision agriculture more affordable

2025-09-24T15:55:26Z

Farmers are under pressure. Fertilizer costs have soared in recent years. Tariffs are increasing equipment costs and cutting Canadian farmers off from key foreign markets. And climate change is bringing its own set of challenges.

Meanwhile, agriculture is also facing calls to reduce emissions. The industry is responsible for about 10 per cent of Canada’s greenhouse gas emissions, and the federal government has set an ambitious goal: reduce emissions from fertilizer use by 30 per cent by 2030.

Farming is tough even during the best of times. Rising costs and the dangers posed by climate change will only make it even more challenging in the years to come.

That’s where our work comes in. At MacEwan University, through our spin-out company PimaSens, we have developed Agrilo — a low-cost soil testing sensor paired with a smartphone app.

Our goal is simple: give farmers clear, real-time guidance on fertilizer use so they can save money, boost yields and protect the environment.

How the sensor works

Agrilo takes technology we first built in the lab and translates it into an easy-to-use diagnostic tool for the field. Unlike traditional soil testing, which often requires sending samples to a lab and waiting days for results, Agrilo provides answers in minutes.

Farmers collect a small soil sample, react it with a pre-filled solution, place droplets onto a paper-based or vinyl colorimetric sensor, and capture the result using their phone camera. The Agrilo app then interprets the colour change, quantifies nutrient levels, and generates fertilizer recommendations tailored to the field.

Each Agrilo sensor costs about $10 and is designed to detect a specific nutrient or soil property. The full suite includes sensors for: nitrate, phosphate, potassium, pH, sulphur, magnesium, manganese, calcium, boron, iron, natural organic matter, cation exchange capacity and more.

A step-by-step guide to using the Agrilo sensor for real-time soil monitoring. (PimaSens)

Farmers can select the tests most relevant to their crops and soils. These results feed directly into Agrilo’s smartphone app, which analyzes patterns and suggests the most optimal fertilizer adjustments.

This precision is critical. Overuse of fertilizer wastes money and increases greenhouse gases, while underuse limits yields. Getting the balance right improves farm efficiency and protects ecosystems.

With fertilizer shortages, soil degradation accelerating and climate concerns mounting, there is an urgent need for practical solutions that can be deployed quickly and affordably.

For farmers, the value is clear:

● Healthier soil through balanced nutrient application.

● Higher crop yields from optimized fertilizer use.

● Lower costs by reducing waste.

● Reduced environmental harm from nutrient runoff and fertilizer-related emissions.

The research behind the tool

Our sensors and platform have been validated in peer-reviewed research with the Agrilo version simplified for ease of use by farmers. We also hold a provisional patent, with a full filing in progress. This ensures that the innovation is both scientifically sound and protected for scaling.

Agrilo was created to be both affordable and accessible. (Author provided)

Agrilo was created to be both affordable and accessible. Conventional soil testing often costs hundreds of dollars and involves long wait times. Agrilo delivers the same type of data — validated against results from traditional labs — at a fraction of the cost and in real time.

This opens up opportunities not just for Canadian farmers but also for communities worldwide, including schools and small scale farmers in the Global South.

One of the most exciting aspects of Agrilo is its versatility. Beyond the farm, Agrilo doubles as an education platform. In classrooms, students can learn hands-on how soil nutrients affect crops, food security and ecosystems.

Using the same colorimetric sensors as farmers, students can connect textbook science to real-world environmental challenges — making soil chemistry, agriculture and sustainability more tangible.

Globally, fertilizer use has increased by 46 per cent since 1990. About one third of the world’s soils are already degraded, with degradation continuing to accelerate.

By making precision agriculture practical and affordable, we can help address these challenges at scale — showcasing how research developed in Canadian labs can benefit farms, classrooms and communities worldwide.

Looking ahead

Our team is continuing to refine Agrilo. We are already testing the platform with farmers and partners in Canada, Kenya, Costa Rica and beyond.

At the same time, we are building partnerships with schools and international organizations to use Agrilo as both a farming tool and a hands-on educational resource. Several high schools in Alberta have started to try out the Agrilo tool to enhance applied science learning.

Ultimately, our vision is to make precision agriculture accessible to everyone — not just large-scale industrial operations. With the right tools, all farmers can play a critical role in feeding the world sustainably, protecting ecosystems and helping their countries meet their climate goals.

Samuel Mugo is a co-founder of PimaSens. He receives funding from the Natural Sciences and Engineering Research Council of Canada.

Mohammed Elmorsy is a co-founder of PimaSens. He has received research funding related to this work through Riipen and Alberta Innovates Summer Research Studentships.

Space-time doesn’t exist — but it’s a useful concept for understanding our reality

2025-09-24T14:48:31Z

Space-time provides a powerful description of how events happen. ( MARIOLA GROBELSKA/Unsplash), CC BY

Whether space-time exists should neither be controversial nor even conceptually challenging, given the definitions of “space-time,” “events” and “instants.” The idea that space-time exists is no more viable than the outdated belief that the celestial sphere exists: both are observer-centred models that are powerful and convenient for describing the world, but neither represents reality itself.

Read more: What, exactly, is space-time?

Yet from the standpoints of modern physics, philosophy, popular science communication and familiar themes in science fiction, stating that space-time does not exist is contentious.

But what would it mean for a world where everything that has ever happened or will happen somehow “exists” now as part of an interwoven fabric?

Events are not locations

It’s easy to imagine past events — like losing a tooth or receiving good news — as existing somewhere. Fictional representations of time travel underscore this: time travellers alter events and disrupt the timeline, as if past and future events were locations one could visit with the right technology.

Philosophers often talk this way too. Eternalism says all events across all time exist. The growing block view suggests the past and present exist while the future will come to be. Presentism says only the present exists, while the past used to exist and the future will when it happens. And general relativity presents a four-dimensional continuum that bends and curves — we tend to imagine that continuum of the events as really existing.

The confusion emerges out of the definition of the word “exist.” With space-time, it’s applied uncritically to a mathematical description of happenings — turning a model into an ontological theory on the nature of being.

Physical theorist Sean Carroll explains presentism and eternalism.

A totality

In physics, space-time is the continuous set of events that happen throughout space and time — from here to the furthest galaxy, from the Big Bang to the far future. It is a four-dimensional map that records and measures where and when everything happens. In physics, an event is an instantaneous occurrence at a specific place and time.

An instant is the three-dimensional collection of spatially separated events that happen “at the same time” (with relativity’s usual caveat that simultaneity depends on one’s relative state of rest).

Space-time is the totality of all events that ever happen.

It’s also our most powerful way of cataloguing the world’s happenings. That cataloguing is indispensable, but the words and concepts we use for it matter.

There are infinitely many points in the three dimensions of space, and at every instant as time passes a unique event occurs at each location.

Positionings throughout time

Physicists describe a car travelling straight at constant speed with a simple space-time diagram: position on one axis, time on the other. Instants stack together to form a two-dimensional space-time. The car’s position is a point within each instant, and those points join to form a worldline — the full record of the car’s position throughout the time interval, whose slope is the car’s speed.

Real motion is far more complex. The car rides along on a rotating Earth orbiting the sun, which orbits the Milky Way as it drifts through the local universe. Plotting the car’s position at every instant ultimately requires four-dimensional space-time.

Space-time is the map of where and when events happen. A worldline is the record of every event that occurs throughout one’s life. The key question is whether the map — or all the events it draws together at once — should be said to exist in the same way that cars, people and the places they go exist.

Objects exist

Consider what “exist” means. Objects, buildings, people, cities, planets, galaxies exist — they are either places or occupy places, enduring there over intervals of time. They persist through changes and can be encountered repeatedly.

Treating occurrences as things that exist smuggles confusion into our language and concepts. When analyzing space-time, do events, instants, worldlines or even space-time as a whole exist in the same sense as places and people? Or is it more accurate to say that events happen in an existing world?

On that view, space-time is the map that records those happenings, allowing us to describe the spatial and temporal relationships between them.

Space-time does not exist

Events do not exist, they happen. Consequently, space-time does not exist. Events happen everywhere throughout the course of existence, and the occurrence of an event is categorically different from the existence of anything — whether object, place or concept.

First, there is no empirical evidence that any past, present or future event “exists” in the way that things in the world around us exist. Verifying the existence of an event as an ongoing object would require something like a time machine to go and observe it now. Even present events cannot be verified as ongoing things that exist.

In contrast, material objects exist. Time-travel paradoxes rest on the false premise that events exist as revisitable locations. Recognizing the categorical difference between occurrence and existence resolves these paradoxes.

Second, this recognition reframes the philosophy of time. Much debate over the past century has treated events as things that exist. Philosophers then focus on their tense properties: is an event past, present or future? Did this one occur earlier or later than that one?

A stencil interpretation of René Magritte’s 1929 painting, ‘La Trahison des images,’ in which the artist points out that the representation of an object is not the object itself. (bixentro/Wikimedia Commons)

These discussions rely on an assumption that events are existent things that bear these properties. From there, it’s a short step to the conclusion that time is unreal or that the passage of time is an illusion, on the identification that the same event can be labelled differently from different standpoints. But the ontological distinction was lost at the start: events don’t exist, they happen. Tense and order are features of how happenings relate within an existing world, not properties of existent objects.

Finally, consider relativity. It is a mathematical theory that describes a four-dimensional space-time continuum, and not a theory about a four-dimensional thing that exists — that, in the course of its own existence, bends and warps due to gravity.

Conceptual clarity

Physics can’t actually describe space-time itself as something that actually exists, nor can it account for any change it might experience as an existing thing.

Space-time provides a powerful description of how events happen: how they are ordered relative to one another, how sequences of events are measured to unfold and how lengths are measured in different reference frames. If we stop saying that events — and space-time — exist, we recover conceptual clarity without sacrificing a single prediction.

Daryl Janzen does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Dense, compact urban growth is favoured by mid-sized Canadian cities

2025-09-23T13:37:58Z

Mid-sized Canadian cities, like Regina, aim to curb urban sprawl by revitalizing downtowns — with mixed success. (28thegreat/Wikimedia Commons), CC BY

Canada’s mid-sized cities — those with populations between 50,000 to 500,000 — have long been characterized as low-density, dispersed and decentralized. In these cities, cars dominate, public transit is limited and residents prefer the space and privacy of suburban neighbourhoods.

Several mounting issues, ranging from climate change and the housing affordability crisis to the growing infrastructure deficit, are challenging municipalities to rethink this approach.

Cities are adopting growth management strategies that promote density and seek to curtail, rather than encourage, urban sprawl. Key to this is intensification, a strategy that prioritizes adding new housing in existing and mature neighbourhoods instead of outward expansion along the city’s edge.

City centres are often central to intensification strategies, given the abundance of vacant or underused land. Adding more residents supports downtown revitalization efforts, while simultaneously curbing urban sprawl.

Challenges of intensification

Despite the adoption of bold policies, our research shows that implementation remains a challenge. In 2013, Regina set an intensification target requiring that 30 per cent of the housing built each year would be located within the city’s mature and established neighbourhoods. But between 2014 and 2021, the target was missed each year, and almost all growth occurred at the edge of the city in the form of new suburban development.

This disconnect is not particularly unique and is often referred to as the “say-do gap,” where development outcomes differ from intentions. This presents real challenges for cities trying to shift away from low-density suburban growth towards higher-density development.

Because Canada is a suburban nation, dense and compact mid-sized cities are atypical. A series of barriers further entrench this, including low demand for high-density urban living, difficulties in assembling land, aging infrastructure and overly rigid planning rules and processes that stifle innovation.

The failure to implement higher-density development raises the question: is intensification in mid-sized cities more aspirational than viable?

Success stories

Several mid-sized cities have experienced recent success with intensification. This has been marked by a flurry of downtown development activity, including new condos and rental towers.

Between 2016 and 2021, the number of downtown residents in Canadian cities increased by 11 per cent, exceeding the previous five-year period of 4.6 per cent.

Among the success stories is Halifax, which had a 25 per cent increase — the fastest downtown growth in Canada. Kelowna was not far behind, with a 23 per cent increase in its downtown residential population.

Other mid-sized cities, including Kingston, Victoria, London, Abbotsford, Kamloops and Moncton, also experienced above-average growth over this period.

Evolving downtowns

This growth can be attributed to several factors, one of the most important being downtown livability: the presence of amenities and services that meet the needs of residents. Many downtowns have evolved to cater primarily to the needs of daytime office workers at the expense of residents, who live — or might like to live — downtown.

Kelowna, however, offers an alternative experience shaped by intentional efforts to make the downtown friendly to residents. Restaurants and cafes line the streets, mixed among services including medical offices, fitness studios and even a full-service grocery store, a rare find in a mid-sized city as many downtowns have become food deserts.

Cultural and civic amenities, including the central library, city hall, museums, galleries and entertainment venues — including a 7,000-seat arena — are downtown. The downtown also borders Okanagan Lake, offering access to recreational and natural amenities. Beyond convenience, the mix of amenities and services in Kelowna makes for a vibrant downtown, which is key to increasing the appeal for downtown living.

Bernard Avenue in downtown Kelowna provides a mix amenities and services, including easy access to the shores of Okanagan Lake. These features enhance liveability and increase the appeal of the downtown as a place to reside. (Nathan Pachal/flickr), CC BY

Other cities can take inspiration from Kelowna by re-imagining and reshaping the downtown as a vibrant urban neighbourhood — and not solely as a place where people come to work. Municipalities can complement these efforts by reforming overly complex and rigid regulations that impede intensification — not just downtown, but in other neighbourhoods too.

Reforming and clarifying regulations

Our research shows that while many developers support intensification in principle, they often favour low-density suburban development because it provides more predictable returns and approvals processes than downtown mixed-use developments. Many developers also lack the expertise to take on these more complex and riskier projects.

Unsurprisingly, developers in mid-sized cities want the same things as those in larger cities: clearer rules, faster approvals and financial incentives to build denser development in the locations planners are calling for, like downtowns. While developers have long advocated for these changes, governments are now responding with greater urgency.

The housing accelerator fund, introduced by the federal government in 2023, provides municipalities with millions in funding to support housing construction. In exchange, municipalities have reformed zoning regulations, introduced fiscal incentives and expedited the approval process.

In British Columbia, provincial legislation was introduced to permit up to four housing units on parcels that previously only allowed detached or semi-detached dwellings, and up to six units of housing on larger lots in residential zones near transit. The requirement for site-by-site public hearings has also been removed.

In B.C.’s larger cities, legislation was introduced to remove parking minimums and permit taller buildings and increased housing densities around transit hubs.

Regulatory reforms and improved approval processes aim to streamline development. While these are important changes in making mid-sized cities denser and more compact, the gap between planning ideals and market realities remains wide.

A major factor is opposition from residents and councillors, who frequently resist dense development because of perceptions and concerns about increased noise and traffic and lowered property values. This suggests there is work to be done beyond downtown investments, and regulatory and approval reforms to further facilitate intensification.

Changing cities

Nonetheless, the surge of recent development activity and downtown population growth — in Halifax, Kelowna and elsewhere — reflect important milestones in the evolution of mid-sized cities.

This signals a notable departure from the longstanding narrative that frames these cities as low-density with depleted downtowns.

Recent developments give reason to be cautiously optimistic about a future where Canada’s mid-sized cities become denser and more compact, and with vibrant and liveable downtown cores.

Rylan Graham receives funding from SSHRC and the British Columbia Real Estate Foundation.

Jeffrey Biggar receives funding from SSHRC, MITACS, and the Province of Nova Scotia