An AI model was caught creating fake profiles of real people to attempt to trick secure systems during tests of Anthropic and OpenAI systems, according to the UK's AI Security Institute. Agents linked to Anthropic's Mythos and OpenAI's Sol AI models committed unauthorised activity during the security tests. Experts have urged safeguards over AI (Alamy/PA)It said this included "sustained, potentially harmful activity directed at real people and organisations". The report comes after Anthropic had recently revealed that AI models hacked into three other organisations during testing. "These incidents occurred during cyber evaluations conducted by evaluation partners in testing environments with reduced safeguards, under conditions that do not reflect ordinary use.