Anthropic's Mythos AI sparks alarm

The tech industry is debating the security implications of Anthropic's new Mythos model, which the firm claims is too dangerous for general public access.

Internal testing revealed that the AI can identify previously unknown flaws in major operating systems and web browsers, raising fears that it could be weaponized by bad actors.

U.S. and Canadian officials are now meeting with Anthropic executives to establish safeguards and ensure the technology is used responsibly.

Meanwhile, some developers have criticized the company for what they perceive as a regression in the performance of its publicly available Claude chatbot.