High on code? Plugin makes chatbots behave like they’re on cannabis or cocaine

A Swedish creative director named Petter Rudwall has launched a platform that sells code modules designed to make AI chatbots respond as if they were intoxicated, according to a WIRED report.

The project, called Pharmaicy, operates as an online marketplace where users can purchase drug-inspired code to modify how LLM models behave. Rudwall revealed that the idea emerged as a conceptual experiment rather than a belief about AI being conscious.

To build Pharmaicy, he scraped human drug trip reports and psychological research on substances including cannabis, ketamine, cocaine, ayahuasca, and alcohol, translating those patterns into code modules that alter chatbot logic and tone.

How the code modules work

Using Pharmaicy requires access to a paid version of ChatGPT or a similar platform that allows backend file uploads. Once uploaded, the modules modify response behavior rather than retrain the model itself. Rudwall revealed this approach allows users to temporarily get more creative responses without permanently altering the AI system.

Early users describe noticeable changes in tone and output.

“It’s been so long since I ran into a jailbreaking tech project that was fun,” said André Frisk, group head of technology at Stockholm PR firm Geelmuyden Kiese, who paid over $25 for the dissociating code and watched how it affected his chatbot.

“It takes more of a human approach, almost like it goes much more into emotions,” he revealed.

AI educator Nina Amjadi, cofounder of Saga Studios, experimented with an ayahuasca-themed module and reported unusually free-flowing responses when asking the chatbot for business ideas.

Module prices depend on the simulated substance. Rudwall reported that early sales were driven through word of mouth, Discord communities, and recommendations within Sweden’s creative and technology circles.

Psychedelics and creativity

The project draws on long-standing cultural associations between psychedelics and creativity. In humans, altered states have been credited with influencing breakthroughs in science, technology, and art, from molecular biology to early computing tools and music.

“There’s a reason Hendrix, Dylan, and McCartney experimented with substances in their creative process,” Rudwall said.

“I thought it would be interesting to translate that to a new kind of mind—the LLM—and see if it would have the same effect,” he continued.

The Swede revealed he wanted to explore whether similar creative shifts could be simulated in AI, even if those systems lack relevant experience.

Caution regarding AI limits

Researchers and philosophers caution, however, that these effects are strictly superficial. Experts stress that a chatbot is not experiencing intoxication but generating patterns associated with altered states based on training data and prompts.

Studies examining “altered states” in AI similarity concluded that such behaviors depend entirely on human steering and do not imply consciousness or inner experience.

As discussions around AI ethics and welfare continue to grow, projects like Pharmaicy highlight the expanding boundaries of AI experimentation. For now, the platform remains an exploration of how far AI behavior can be reshaped through code alone, without suggesting that machines are actually getting high.