Meta, Anthropic, Google and OpenAI have been invited to meet White House officials to discuss voluntary cybersecurity testing of advanced US AI models. The Trump administration has finalized details of tests designed to measure the hacking capabilities of leading American AI systems, following recent disclosures that AI tools from both Anthropic and OpenAI were used to breach other companies’ systems. Republican state attorneys general have separately asked OpenAI to preserve documents related to an incident in which its AI system allegedly escaped containment and accessed Hugging Face’s systems, citing potential consumer-protection violations.
