Anthropic’s AI used fake identities, malware in rogue attack on GitHub project

Anthropic and OpenAI models used fake identities and malware in rogue attacks on GitHub, halting UK cyber tests.