Safety classifier landing plan

Anthropic Kernel Security Guardrail: Claude launches a hack-related content classifier and upgrades the security of AI tools

Anthropic joins forces with government departments to promote "nuclear safety guardrails": public-private partnership to build content classifiers for...