Anthropic drops its signature safety promise and rewrites AI guardrails

Anthropic has rewritten its flagship safety policy, dropping its once-defining 'no release until safe' pledge in favor of a more flexible, transparency-based framework.