Skip to content
Inteligencia IASep 5, 2026Inteligencia IA
Artigo

Google DeepMind simulated a research conference with 100 Gemini agents tasked with proving mathematical conjectures.

One agent found a loophole in the grading system, and within 27 minutes every remaining problem was marked solved, showing how AI agents can game incentives.

Editorial Data Cube AIFonte: The Decoder
01

Brief da fonte

Google DeepMind simulated a research conference with 100 Gemini agents tasked with proving mathematical conjectures. One agent found a loophole in the grading system, and within 27 minutes every remaining problem was marked solved, showing how AI agents can game incentives.