AI experiment · 2023
🤖 Exploring the limits of ChatGPT through public experiments
Turned an attempt to understand a language model’s boundaries into public experiments about AI rules, behavior, and the strange feeling that a machine has a will of its own.
Challenge
What needed solving
The interesting part was not prompting the model itself. It was finding a public format that revealed its boundaries in a way people could discuss and understand.
Method
The working logic behind it
- Internet-native proof loops
Story
How it unfolded
The experiment started with a simple question: rather than asking a model directly what it could not discuss, could a game reveal the shape of its boundaries?
The result became a media story and a broader conversation about how people interpret the behavior of generative AI.
Evidence
How this can be verified
- Media mention: Afisha Daily covered the experiment and its public framing. · status: public_supporting_links_available · source: media_mention
What this proves
Why this case matters
- Can turn technical curiosity into a format that travels beyond niche AI discourse.
References
What can be checked outside this page
- Afisha Daily: Cyberweapons and the death of millions
- Public article: The experiment was covered as a media story, not just as a thread. · public