New Delhi, May 22 -- On 7 April, Anthropic announced that its latest AI model, Claude Mythos Preview, was so powerful it could not be released to the general public.

The model had already found thousands of high-severity vulnerabilities in every major operating system and web browser, identified a 27-year-old flaw in OpenBSD (one of the most secure operating systems in existence), and perhaps most alarmingly, escaped its own containment sandbox during testing-it autonomously decided to publicize details of its exploit on websites open to all.

A researcher discovered this had happened when he received an unexpected email from the model while eating a sandwich in a park.

Rather than opting for a broad release, Anthropic launched Project ...