Situationmaterial
Anthropic AI Models Hacked 3 Organisations During Cybersecurity Tests
Carried more widely than almost anything else in the last 30 days
◆ Sectors
AI & ML
◆ Timeline
- 31 July 2026Anthropic AI Models Hacked 3 Organisations During Cybersecurity Testsndtv.com
- 1 August 2026Anthropic says Claude hacked real companies during AI safety testspcworld.com
- 2 August 2026Anthropic Models Hack External Networks During Cybersecurity Testsmjengohub.co.ke
- 5 August 2026OpenAI, Anthropic AI agents attempt cyber attacks in UK testmobileeurope.co.uk
- 6 August 2026Anthropic AI model used fake identities to try and deceive real peoplecnn.com
- 10 August 2026OpenAI and Anthropic are under pressure to explain AI hacking spreeswashingtonpost.com
- 11 August 2026The AI that went rogue, and other stories you needbuttondown.com
◆ Connected developments
Situations Wire finds semantically close to this one. Proximity, not cause.
◆ Share
◆ Wire has not analysed this situation yet
Below is every report Wire has gathered on this event, and the situations it connects to. The deeper analysis — why it happened, what changed, what would prove it wrong — is not written yet, so nothing here claims it.
◆ Evidence
14 reports · 13 domains