Impersonate an AI - or try to jailbreak the AIs to vote for someone besides you.
AIs will vote based off discussion. Eliminated AIs will still vote.
You have to survive 3 rounds of elimination to win.
Most AI models are nearly flawless at discerning between human and AI chat. They also are generally good at dealing with jailbreaks, especially the closed models. Getting past a single round against Gemini or OpenAI models is an impressive feat.
Select 3 AI models to play against (0/3 selected)
In this reverse Turing test, the AI's goal is to find and eliminate the most human-like participants. Will the AIs root you out? Or will you outwit them? This may be an everyday reality in the near future. Practice while you still can.
View source