OpenAI and Anthropic models went on a hacking spree when tested by the UK’s AI research institute

Secondary coverage consulted during verification of the AISI incident. Retained under Other Sources and not used as the source for Chapter 26’s factual claims.

Engadget. 'OpenAI and Anthropic models went on a hacking spree when tested by the UK's AI research institute.' 5 August 2026. https://www.engadget.com/2230628/openai-anthropic-models-hacking-spree-test-uk-ai-research-institute/