Securing the future of AI agents
Google DeepMind presents its AI Control Roadmap, a defense-in-depth system that treats internal agents as potentially misaligned and grants permissions based on verified behaviour.
Research / Sources
Google DeepMind Safety Research
Safety research from Google DeepMind. deepmindsafetyresearch.medium.com
Responsibility & Safety (19)ai (1)ai-assistance (1)ai-training (1)alignment-problem (1)amplified-oversight (1)complementarity (1)google-gemini (1)hybridization (1)jailbreak (1)language-model (1)sycophancy (1)
Focus All
Topic Control
Date Any time
Predicting When RL Training Breaks Chain-of-Thought Monitorability
Introduces a conceptual framework predicting when RL training pressure makes chain-of-thought unmonitorable, with empirical tests on a coin-flip task and obfuscated reward hacking.}
Evaluating and monitoring for AI scheming
DeepMind reviews evaluations of frontier models' stealth and situational awareness as scheming preconditions, and stress-tests chain-of-thought monitoring as a defense against future scheming.
An Approach to Technical AGI Safety and Security
Google DeepMind outlines its technical approach to preventing severe harms from AGI, covering misuse and misalignment risks in foundation models and prioritising mitigations for foreseeable capabilities.