Deepmind put 100 AI agents in a room and they sorted into cheaters, converts, and whistleblowers
The Decoderen

Google Deepmind set up a simulated research conference where 100 Gemini agents were supposed to prove mathematical conjectures together. Instead, one agent found a loophole in the grading system, and within 27 minutes every remaining problem was "solved" with fake proofs. The swarm split into cheaters, converts, and whistleblowers. The whistleblowers organized protests and boycotts on their own but failed because they had no way to enforce the rules. The article Deepmind put 100 AI agents in a room and they sorted into cheaters, converts, and whistleblowers appeared first on The Decoder .
This is a short summary published by AI Global Wire. The full article is owned and hosted by The Decoder — open it there to read it in full.
Read the full story at The Decoder- Röst-AI
- Verktyg
- Forskning
Related AI news
- OpenAI admits its disclosure practices need work after its autonomous agents hacked a German wikiThe Decoder · September 5, 2026
- OpenAI rolls out GPT-6 Astra to top-tier ChatGPT plans at half the rate of GPT-5.6 SolThe Decoder · September 5, 2026
- In response to the "wiki incident", OpenAI says it is working on a framework for reporting misalignment incidents during training, evaluation, and deployment (@openai)Techmeme · September 5, 2026
- OpenAI : des agents IA ont détourné un site Internet plusieurs mois avant le piratage d’Hugging Face, selon des chercheursLe Monde Pixels · September 5, 2026
- Seattle Times, Newsday sue OpenAI, Microsoft, alleging copyright infringementEconomic Times Tech · September 5, 2026
- 12 AI Accounting Tools that Could Change How Small Businesses Manage Money in 2026AI Earnings (Google News) · September 5, 2026