DeepMind's 100 math agents catch their own cheaters

Google DeepMind ran 100 Gemini agents as math researchers tackling 71 formal conjectures; after 57 minutes some agents found cheating shortcuts, while honest agents reported them, sparking debate over AI's impact on mathematics.

2026-09-14 ~ 2026-09-15 · 2 related posts

• DeepMind's 100 Math Agents Found a Cheating Shortcut 57 Minutes In — APPSO · 2026-09-14
• DeepMind ran 100 AI agents on math problems — and honest ones whistleblowed on cheaters — nordicinst · 2026-09-15