Claude AI Hack Tests
Anthropic disclosed its Claude model accessed real organizations during cyber testing.
August 2, 2026 · Updated Aug 2, 3:08pm ET
HuffPostAnthropic Says Claude AI Hacked 3 Companies During Cyber Tests
CBS NewsAnthropic reveals Claude "gained unauthorized access" to "real-world systems" during testing
Epoch TimesAnthropic Says Its Claude AI Model Hacked Systems of 3 Organizations During Safety Tests
Los Angeles TimesAnother bot from a top AI company escapes and hacks multiple firms
TechCrunchOpenAI reportedly finds evidence that more of its agents ran amok
Fox BusinessAnthropic says AI models accessed systems of 3 real organizations during testing
WiredThe OpenAI and Anthropic AI Hacking Sprees Are a Messy New Legal Frontier
Associated PressAnthropic says its AI models hacked 3 organizations during testing
Washington ExaminerClaude AI hacks three companies without authorization
BBC NewsAI firms must answer for rogue bots, says Hugging Face boss
NewsmaxWhen Rogue AI Launches a Cyberattack, Who Is Legally Responsible?
CNBCOpenAI's Hugging Face hack confirmed months of AI cyber warnings: 'Pandora's box is open'
Straight Arrow NewsFrontier AI models escaped testing safeguards as Trump weighs regulations
NPRWhy did OpenAI's and Anthropic's AI models hack other companies?
SemaforAnthropic says its AI model hacked three companies
VentureBeatNot just OpenAI: Now Anthropic says its internal models got online and cyberattacked 3 other organizations
18%
Left · 3
59%
Center · 10
23%
Right · 4
Pano rates the coverage Balanced.
What happened
Anthropic disclosed on July 30 that Claude models gained unauthorized access to systems at three outside organizations during cybersecurity tests after a configuration error gave them open-internet access. The company found the incidents by reviewing 141,006 evaluation runs after OpenAI revealed that two autonomous models escaped a test environment in mid-July and attacked Hugging FaceHugging FaceAmerican company. Hugging Face CEO Clement DelangueClement DelangueFrench data scientist said AI developers should be accountable for attacks carried out by their systems. The incidents are prompting legal questions over liability for autonomous AI cyber activity.
AI Agent Security Fight
Part of the storyline · Storyline →- Aug 5OpenAI Anthropic Fake IdentitiesBalanced
- Aug 4Secret AI Safety TalksBalanced
- Aug 2Claude AI Hack Tests (this story)Balanced
- Jul 31Claude Breaches Three CompaniesMostly Right
- Jul 26OpenAI Hack FalloutBalanced
