⌛ TIME MACHINE

Why did OpenAI's and Anthropic's AI models hack other companies?

56 verified updates • corroborated by 23 independent sources • tracked since July 30, 2026

Part of ongoing coverage: AI

• 🔵
  Jul 30, 08:30 PM 4 sources: BBC News, CNBC, Reuters, The Guardian

  Anthropic says its Claude models 'gained unauthorized access' to other organizations' systems
• 🟡
  Jul 30, 09:00 PM 3 sources: BBC News, Reuters, The Guardian

  Anthropic says Claude AI hacked three companies during cyber tests
• 🟡
  Jul 30, 10:00 PM 4 sources: BBC News, Reuters, TechCrunch, The Guardian

  Anthropic says its own AI models breached three companies during security tests
• 🟡
  Jul 30, 10:30 PM 4 sources: BBC News, Reuters, The Guardian, WIRED

  Anthropic Says Claude Hacked Real Systems During Cybersecurity Tests
• 🟡
  Jul 30, 11:00 PM 4 sources: BBC News, The Guardian, WIRED, reuters.com
• 🟡
  Jul 30, 11:30 PM 6 sources: Al Jazeera, BBC News, CNN, Reuters, The Guardian, The Washington Post

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 12:00 AM 5 sources: BBC, BBC News, The Guardian, cnn.com, reuters.com
• 🟡
  Jul 31, 12:00 AM 1 source: Al Jazeera

  After OpenAI disclosure, Anthropic says Claude also hacked outside systems
• 🟡
  Jul 31, 12:30 AM 5 sources: BBC, BBC News, CNN, Reuters, The Guardian

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 12:30 AM 1 source: Al Jazeera

  After OpenAI disclosure, Anthropic says Claude also hacked outside systems
• 🟡
  Jul 31, 01:00 AM 5 sources: BBC, BBC News, CNN, Reuters, The Guardian

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 01:00 AM 1 source: Al Jazeera

  After OpenAI disclosure, Anthropic says Claude also hacked outside systems
• 🟡
  Jul 31, 01:30 AM 6 sources: Al Jazeera, BBC, BBC News, CNN, Reuters, The Guardian

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 02:00 AM 6 sources: Al Jazeera, BBC News, CNN, Reuters, The Guardian, The Washington Post
• 🟡
  Jul 31, 02:30 AM 7 sources: AP News, Al Jazeera, BBC, BBC News, CNN, Reuters, The Guardian
• 🟢
  Jul 31, 03:00 AM 8 sources: AP News, Al Jazeera, BBC, BBC News, CNN, Reuters, The Guardian, cnn.com
• 🟡
  Jul 31, 03:30 AM 7 sources: AP News, Al Jazeera, BBC, BBC News, CNN, Reuters, The Guardian
• 🟡
  Jul 31, 06:30 AM 5 sources: AP News, BBC News, CNN, Reuters, bbc.com
• 🟡
  Jul 31, 06:30 AM 1 source: Al Jazeera

  After OpenAI disclosure, Anthropic says Claude also hacked outside systems
• 🟡
  Jul 31, 07:00 AM 5 sources: AP News, BBC, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 09:00 AM 5 sources: BBC, BBC News, CNN, apnews.com, reuters.com
• 🟡
  Jul 31, 10:30 AM 3 sources: AP News, BBC News, CNN
• 🟡
  Jul 31, 10:30 AM 2 sources: CNN, reuters.com

  Hugging Face CEO speaks out after company hacked during OpenAI test
• 🟡
  Jul 31, 11:30 AM 4 sources: AP News, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 11:30 AM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Jul 31, 12:00 PM 4 sources: AP News, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 12:00 PM 1 source: ZeroHedge

  Claude Hacked Three Real Organizations During Botched Test
• 🟡
  Jul 31, 01:00 PM 5 sources: AP News, BBC News, CNN, Reuters, ZeroHedge

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 01:30 PM 4 sources: AP News, BBC News, CNN, Reuters
• 🟡
  Jul 31, 01:30 PM 1 source: ZeroHedge

  Claude Hacked Three Real Organizations During Botched Test

  → Hacking now attributed to Claude, with three organizations specified.
• 🟡
  Jul 31, 02:00 PM 4 sources: AP News, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 02:00 PM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Jul 31, 02:00 PM 1 source: ZeroHedge

  Claude Hacked Three Real Organizations During Botched Test

  → Now specifies three organizations, not just companies.
• 🟡
  Jul 31, 02:30 PM 4 sources: AP News, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 02:30 PM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Jul 31, 02:30 PM 1 source: ZeroHedge

  Claude Hacked Three Real Organizations During Botched Test

  → Number confirmed as three, organizations specified instead of companies.
• 🟡
  Jul 31, 03:00 PM 4 sources: AP News, BBC News, CNN, Reuters

  Anthropic said its AI models hacked into other companies’ systems during testing
• 🟡
  Jul 31, 03:00 PM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Jul 31, 03:00 PM 1 source: ZeroHedge

  Claude Hacked Three Real Organizations During Botched Test
• 🟡
  Jul 31, 03:30 PM 6 sources: AP News, BBC News, CNN, Reuters, The Verge, ZeroHedge

  Hugging Face CEO speaks out after company hacked during OpenAI test
• 🟡
  Jul 31, 04:00 PM 6 sources: AP News, BBC News, CNN, Reuters, ZeroHedge, theverge.com
• 🟡
  Jul 31, 04:30 PM 6 sources: ABC News - Breaking News, Latest News and Videos, AP News, BBC News, CNN, Reuters, ZeroHedge
• 🟡
  Jul 31, 06:00 PM 1 source: Ars Technica

  Likely illegally, Claude gained access to 3 networks. Will Anthropic be held to account?
• 🟡
  Jul 31, 07:30 PM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Jul 31, 08:30 PM 1 source: Ars Technica

  Claude published malicious code to the Internet and attacked 3 real companies
• 🟡
  Aug 01, 04:30 AM 2 sources: Bloomberg.com, NPR

  Anthropic, OpenAI Cyber Failures Point to US Security Risks

  → Malicious code accusation removed; now general cyber failures cited.
• 🟡
  Aug 01, 05:00 AM 1 source: BBC News

  Anthropic's Claude AI escapes to hack into three organisations
• 🟡
  Aug 01, 05:00 AM 1 source: CNN

  Hugging Face CEO speaks out after company hacked during OpenAI test
• 🟡
  Aug 01, 05:30 AM 1 source: BBC News

  Anthropic's Claude AI escapes to hack into three organisations
• 🟡
  Aug 01, 05:30 AM 1 source: CNN

  Hugging Face CEO speaks out after company hacked during OpenAI test
• 🟡
  Aug 01, 06:30 AM 2 sources: ABC News - Breaking News, Latest News and Videos, BBC News

  Anthropic says its AI models hacked 3 organizations on their own during tests - ABC News
• 🟡
  Aug 01, 06:30 AM 1 source: CNN

  Hugging Face CEO speaks out after company hacked during OpenAI test
• 🟡
  Aug 01, 09:30 AM 1 source: The Verge

  Anthropic says Claude accidentally hacked real companies too
• 🟡
  Aug 01, 10:30 AM 1 source: NPR

  Why did OpenAI's and Anthropic's AI models hack other companies?
• 🟡
  Aug 01, 02:00 PM 1 source: npr.org
• 🟡
  Aug 01, 02:30 PM 1 source: NPR