Mo Bitar discusses Anthropic’s governance, external model evaluation and the wider AI safety community. He argues that overlapping personal and institutional relationships create questions about the independence of oversightGovernance sets responsibilities, rules and oversight for decisions about how AI systems are used..
Mo Bitar traces relationships involving the company’s trust, outside evaluation and government safety work, drawing on documents he describes in the video. These are his interpretations and allegations, not findings independently verified by the video.
He notes that an outside evaluator identified problems in several safety reviewsEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. while agreeing with the company’s overall risk judgments. His discussion ends with questions about shareholder influence and governance.
Watch on YouTube
