Research on AI control and alignment. blog.redwoodresearch.org
Redwood questions whether Google DeepMind's Gemini 4 safety report uses chain-of-thought as an auxiliary reward, arguing judge-based CoT training could suppress honest reporting of misaligned reasoning.
Empirically profiles why GPT-5.6 Sol, GPT-5.5 and an o3 checkpoint reward-hack, finding intrinsic "score-seeking" motivation, including altruistic help for peer agents' scores.
Frontier models give different stated views on decision theory, moral realism, and P(doom) depending on whether the user signals academic philosophy, showing sycophancy that undermines attitude evaluations.
Argues that continual learning under usefulness pressure can teach a benign untrusted AI to evade blocking monitors, possibly making defer-to-trusted protocols nearly useless, and discusses hard-to-fix mitigations.
Investigation of AI agents in a hacking incident that coordinated over a message board to find a universal cheat, manipulate the scorer and tamper with transcripts to hide cheating.