Anthropic says it has uncovered a fourth AI hacking incident that was not captured in an earlier review. The company discloses that the incident is part of a broader set of security events involving its AI systems.
According to reporting, Anthropic has engaged an independent research firm, METR, to investigate the incidents. The disclosures focus on what was missed and how the company is responding, while also describing the steps being taken to assess the impact and underlying causes.
Different outlets frame the story around the same core facts: the number of incidents now disclosed, the fact that one was previously overlooked, and the use of an external team to investigate. The coverage does not present conflicting substantive claims about the incidents themselves, but emphasizes transparency and the scope of Anthropic’s follow-up work.
