Anthropic's 154-page official report reveals that Yemen-based Houthi groups and cybercriminals attempted to misuse its 'Claude' AI model to develop bio-weapons, guided missiles, and rocket navigation software. Fortunately, Anthropic's security systems successfully thwarted these malicious attempts in a timely manner.