Anthropic ‘dumps’ its core safety promise on the day its CEO faced angry Pentagon chief

Anthropic is shifting from self-imposed AI safety guardrails to a nonbinding framework, citing an “anti-regulatory political climate.” This change follows a meeting where the Pentagon reportedly issued an ultimatum regarding AI safeguards. The company’s updated Responsible Scaling Policy now separates internal plans from industry-wide recommendations and introduces a Frontier Safety Roadmap and Risk Reports for greater transparency.