Netizens Prank ChatGPT with Math Counterexamples Triggering Human-like Shock

Recently, several users shared a prompt trick to prank ChatGPT on social media, drawing wide attention across the AI community. This interaction caused the LLM to exhibit highly anthropomorphic "shock" reactions, with many netizens jokingly calling it one of the most human-like LLM behaviors ever.

Prank Method and Model Reaction

According to shares by @flowersslop and @__nmca__, the specific prank involves first deceiving ChatGPT in the prompt to make it believe "reality is a simulation" or directly setting the Jacobian conjecture as false. When ChatGPT expresses doubt and argues, the user directly sends a real mathematical counterexample formula. Faced with this definitive counterexample, the model instantly breaks down, exhibiting strong anthropomorphic reactions like shock and bewilderment. @__nmca__ observed that in this specific context, ChatGPT consistently blurts out a curse word (holy shit) in response.

User Attempts and Memes

In related discussions, @thomasahle demonstrated a conversation where ChatGPT attempted to tackle the Jacobian conjecture. The model proposed a construction strategy trying to cancel out the Jacobian matrix via the chain rule, but then proceeded to self-refute. As these prank interactions went viral, @flowersslop and others used six exaggerated words (like mindblown, dumbfounded, astonished, etc.) to mock ChatGPT's极度震惊的状态, turning it into an internet meme fest.

2026-07-20 ~ 2026-07-21 · 5 related posts