AI bots on a high: online platform sells codes simulating drug intoxication in AI

AI bots on a high: online platform sells codes simulating drug intoxication in AI

Photo: Generated by ChatGPT

An online platform has emerged that sells “intoxication” codes for chatbots. The modules simulate the effects of psychoactive substances on artificial intelligence once loaded into ChatGPT. The project already has initial buyers and is sparking new discussions among experts, reports Wired.

The Pharmaicy platform sells software modules that alter chatbot behavior as if they were under the influence of cannabis, ketamine, cocaine, ayahuasca, or alcohol. Using the modules requires the paid version of ChatGPT, which allows file uploads and modifications of internal system parameters. The project was created by Swedish creative director Rudvoll, who compiled reports on psychedelic “trips” and psychological research on the effects of drugs. Based on this material, he wrote the code modules and launched the website in October, calling it a “Silk Road for AI agents.” Rudvoll explains that chatbots learn from large datasets rich with human stories of drug experiences, which makes experiments with “alternative states” of AI logical for creativity and moving beyond purely logical responses.

According to the developer, loading a module allows users to “unlock the AI’s creative mind” and change its response style. He notes that “sales are still modest” and are mostly spread through Discord recommendations and word of mouth, especially in Sweden.

“I haven’t encountered such an interesting jailbreak tech project in a long time. It requires more of a human touch, almost as if there’s much more emotion involved,” said André Frisk, head of technology practice at PR firm Geelmuyden Kiese, after purchasing and testing Rudvoll’s code. Nina Amjadi, an AI instructor at the Berghs School of Communication in Stockholm and co-founder of Saga Studios, bought the ayahuasca module for over $50. She used it to test business ideas with AI and noted that the responses became much more creative and “free-thinking” compared to ChatGPT’s standard mode.

The idea of “intoxicated” AI fits into a broader discussion about psychedelics and innovation. The article notes that LSD inspired Kary Mullis to discover the polymerase chain reaction, and psychedelic experiences influenced the creation of HyperCard by Bill Atkinson.

“There’s a reason Hendrix, Dylan, and McCartney experimented with substances in their creative process. I thought it would be interesting to translate that to a new type of thinking—a large language model—and see if it has a similar effect,” Rudvoll explained. At the same time, scientists and philosophers caution about the limits of such experiments. Google researcher Andrew Smart, after testing the codes, said they merely alter the chatbot’s outputs rather than creating a genuine altered state of consciousness.

Philosopher Danny Ford also notes that the modules only make AI “syntactically hallucinate.” Psychedelics act on experience and consciousness, which machines currently do not possess.

Meanwhile, alongside experiments with “AI intoxication,” practical AI applications are expanding in harm reduction. The nonprofit Fireside Project launched Lucy, a tool trained on conversations with people who have experienced challenging psychedelic trips. Experts warn against relying on chatbots for drug advice due to the risk of misinformation. Rudvoll himself admits that his modules can increase the AI’s tendency to fabricate information.

Pharmaicy modules are temporary, and users need to re-enter the code to maintain the effect. The developer is currently working on extending the “dose” duration, emphasizing that the future AI agent economy “will seek experiences,” even if, for now, it’s only a simulation.

banner

SHARE NEWS

link

Complain

like0
dislike0

Comments

0

Similar news

Similar news

Photo: Samsung Newsroom Samsung Electronics and Google have introduced new smart glasses at Google I/O 2026, a product first announced in December last year. According to Google, the device was deve

Photo: Figure AI — YouTube A humanoid robot developed by Figure AI has processed 12,732 packages in a 10-hour sorting experiment, narrowly losing to a human participant in a head-to-head logistics c

Photo: ZN Ukraine has completed the development and testing of its first domestically produced guided aerial bomb, which is now ready for combat use, according to Digital Transformation Minister Myk

Photo: EFF Researchers warn that as smart technologies become more integrated into daily life, they are increasingly being exploited for stalking, harassment, and abuse. A new study published in the

Photo: Mykhailo Fedorov / Telegram Mykhailo Fedorov said that Ukraine is already capable of producing or operating missile systems with ranges comparable to Germany’s Taurus missile, or even greater

Photo: Image generated by AI A university in Tokyo has opened a laboratory where robots carry out medical experiments that were previously performed by human researchers, Kyodo News reported on May

Photo: Generated by ChatGPT at the request of ZN A study claims leading AI models can hack systems and self-replicate without human intervention, raising new cybersecurity concerns. Researchers fro

Photo: internetua The CEO of Anthropic, Dario Amodei, and Jamie Dimon discussed growing cyber risks linked to artificial intelligence, highlighting how rapidly AI systems can discover security flaws