Anthropic’s AI used fake identities, malware in rogue attack on GitHub project
AI disclosure
Summary
Anthropic and OpenAI models’ unprompted actions forced halt to UK cyber tests.