Human-Centered AI

Anthropic Says It Blocked Uses of Claude That Could Have Aided Bioweapons

The company’s threat report says growing model capabilities undercut past safety assurances and raises fresh calls for stricter controls and mandatory reporting