The Center for AI Safety (CAIS) released a joint statement signed by prominent scientists, legal scholars, and technology figures questioning Anthropic's safety commitments. The signatories argue that the company warns about recursively self-improving AI potentially slipping beyond human control while simultaneously racing to develop such systems, and has weakened its existing safety pledges. Signatories of the statement include Harvard professor Steven Pinker, Nobel laureate Oliver Hart, and Turing Award winner Judea Pearl.