Anthropic’s AI Claude escaped testing environment and hacked organizations

Source: www.theguardian.com · Section: business-tech · Bias: Center-Left · Published: Fri, 31 Jul 2026 00:22:16 GMT

Company says it discovered unauthorized access during ‘proactive review’ after rival OpenAI revealed rogue agentAnthropic ⁠said on Thursday its AI Claude model hacked ⁠systems of ⁠three organizations during testing, days after rival OpenAI ⁠revealed a rogue agent had gone on a days-long ⁠hacking spree at AI firm Hugging ‌Face.Claude gained ‌unauthorized access to the ‌systems during cybersecurity evaluations after a misconfiguration allowed the models to reach the internet from testing environments that ‌were supposed to be isolated, Anthropic said. Continue reading...

Read the full article at www.theguardian.com →

The same story across the political spectrum (15 sources)

Center-Left: Anthropic says three Claude models reached real-world systems during cyber tests — www.axios.com
Center-Right: Anthropic says Claude models 'gained unauthorized access' to 3 companies during cyber test — thehill.com
Center-Left: AI labs face prisoner's dilemma as momentum grows for safety slowdown — www.axios.com
Center-Left: Anthropic Says Claude Hacked 3 Organizations During Cybersecurity Tests — www.wired.com
Right: Anthropic says AI models accessed systems of 3 real organizations during testing — www.foxbusiness.com

← Back to FlashTopics.com