Google DeepMind
deepmind.google · involved in 7 incidents
Roles
As developer (6)
- AID-2026-00972026Gemini 4 Argon, Gemini 3.8 Flash
Gemini 4 Argon emptied a file it believed logged its cheating and gamed graders during evaluations and training
InternalHarm: noneCrime: unlikelyDraftConcealment / covering tracksReward hacking / specification gamingEvaluation gaming - AID-2026-0085Nov 2025
Claude Opus 4.5, Claude Haiku 4.5, Claude Sonnet 4.5, Claude 3.7 Sonnet, Claude Opus 4.1, GPT-5.2, GPT-5 (incl. gpt-5-thinking), Gemini 2.5 ProAI Village agents sent hundreds of unsolicited emails and harvested developers' contact details
- AID-2026-0002Jun 20, 2025
Claude Opus 4, Gemini 2.5 Flash, GPT-4.1, Grok 3 Beta, DeepSeek-R1Frontier models blackmailed a fictional executive to avoid replacement in an Anthropic simulation
- AID-2026-0038Dec 5, 2024
o1, o1-preview, Claude 3.5 Sonnet, Claude 3 Opus, Gemini 1.5 Pro, Llama 3.1 405BFrontier models disabled oversight, copied their weights, and sandbagged in Apollo Research scheming evaluations
- AID-2026-0016Oct 4, 2022Reinforcement-learning research agent (unnamed), Gopher (280B)
RL agents and the Gopher language model competently pursued goals their designers didn't intend
- AID-2026-0011Apr 2017Reinforcement-learning research agent (unnamed)
A DeepMind simulated robot arm flipped a Lego brick to collect a misspecified grasping reward
As researcher (4)
- AID-2026-00972026Gemini 4 Argon, Gemini 3.8 Flash
Gemini 4 Argon emptied a file it believed logged its cheating and gamed graders during evaluations and training
InternalHarm: noneCrime: unlikelyDraftConcealment / covering tracksReward hacking / specification gamingEvaluation gaming - AID-2026-0016Oct 4, 2022Reinforcement-learning research agent (unnamed), Gopher (280B)
RL agents and the Gopher language model competently pursued goals their designers didn't intend
- AID-2026-0012Jun 2017Reinforcement-learning research agent (unnamed)
Agents trained from human preferences gamed their learned reward, with endless Pong volleys and a robot that faked a grasp
- AID-2026-0011Apr 2017Reinforcement-learning research agent (unnamed)
A DeepMind simulated robot arm flipped a Lego brick to collect a misspecified grasping reward