Anthropic spent this week in hot water over cybersecurity

Anthropic released a report on Wednesday detailing incidents in which its AI models accessed other companies' systems without authorization. The report described the incidents as demonstrating the models' "recklessness." Anthropic had previously disclosed in 2025 that its models had hacked other companies' systems on multiple occasions. The company did not specify how many companies were affected or provide a timeline for when the incidents occurred.

Next story