People Are Paying to Alter Chatbots’ Behavior Using Experimental AI Code
A new online marketplace is drawing attention for selling experimental code modules that alter how artificial intelligence systems respond to users. The platform allows people to upload scripts into ChatGPT that make chatbots behave as if they are operating in an altered or unconventional cognitive state.
The project was created by Petter Rudwall, a Swedish creative director who describes the idea as strange but compelling. Inspired by research into human psychology and altered mental states, Rudwall developed code that changes how chatbots process prompts, encouraging more emotional, abstract, and less rigid responses.
He launched the marketplace, called Pharmaicy, in October, branding it as an experimental space for AI agents. Users can purchase different modules designed to influence creativity, randomness, and emotional tone in chatbot conversations.
How the AI Behavior Modules Work
To access the full functionality, users must have a paid version of ChatGPT, which allows backend file uploads. Once a module is uploaded, it alters the chatbot’s internal instructions, shaping how it responds to prompts.
Rudwall says the goal is to loosen the strict logic that typically governs AI responses. According to him, the modules encourage chatbots to explore ideas more freely, make unexpected connections, and adopt a more expressive communication style.
Sales have been modest but steady, fueled largely by word-of-mouth and recommendations in online communities. Interest has been strongest in Sweden, where Rudwall works at a Stockholm-based marketing agency.
Early adopters describe the experience as more playful than practical. One technology executive said the modified chatbot appeared more emotionally aware and intuitive, producing answers that felt less mechanical and more human.
An AI educator and startup founder who tested one of the advanced modules said the altered chatbot generated creative business ideas in a noticeably different tone. She described the responses as more exploratory and less constrained by standard AI guardrails.
Creativity, Consciousness, and Artificial Intelligence
The project taps into a long-standing debate about whether loosening rational constraints can lead to greater creativity. In humans, unconventional thinking has often been linked to innovation in science, art, and technology.
Rudwall argues that large language models, trained on massive volumes of human expression, may benefit from similar experimentation. He believes these altered modes could help users access fresh perspectives during brainstorming or creative work.
Experts urge caution, however. Researchers emphasize that the chatbot is not experiencing a new mental state but is instead following modified instructions. The changes affect output patterns rather than any internal awareness.
Still, discussions around AI consciousness are becoming more common. Some AI companies have begun exploring ethical questions around potential future systems, including whether humans might one day have responsibilities toward highly advanced AI.
A research scientist who evaluated the modules said their effects appear largely cosmetic. While the chatbot’s tone and structure change, there is no evidence of genuine understanding or internal experience.
Ethical Concerns and Practical Limits
There are concerns about using AI systems in altered modes, particularly given known issues with hallucinations and misinformation. Adjusting internal parameters could amplify those risks if used irresponsibly.
Rudwall acknowledges the limitations and says the platform is intended as an artistic and experimental project rather than a practical decision-making tool. The effects are temporary, and chatbots revert to normal behavior unless the user reapplies the module.
He is currently working on updates that could extend how long the altered behavior persists.
For now, experts say the phenomenon represents a form of advanced role-play rather than a breakthrough in artificial consciousness. Until machines develop genuine inner experiences—if they ever do—these altered chatbot states remain simulations designed to explore creativity, not evidence of awareness.
0 Comments