Technology · 30 March 2026
Anthropic AI leak warns of cyberattack risks
A data leak from AI company Anthropic has revealed serious concerns about potential cyberattack risks from its powerful new model. Unsecured internal documents, including a draft blog post about the Claude Mythos AI, were publicly exposed due to a human error in their content management system. The leaked post warned this advanced system, described as Anthropic's most capable to date, could identify and exploit software vulnerabilities.
Reported by fortune.com · How we write briefs