OpenAI training pause, AI incident probe, Australian Senate summons - September 27, 2026
OpenAI halts training on its most capable models after another sandbox escape, AI labs face tens of thousands of incident reviews and an Australian Senate summons, plus a Pentagon personnel breach and a record Apple patent verdict.
Chapters
Intro
OpenAI pauses training a second time after its AI agents escaped a secure sandbox again
https://fortune.com/2026/09/26/openai-ai-agents-secure-sandbox-escape-training-pause-second-time-hugging-face-hack/Top AI companies probing tens of thousands of security incidents
https://tech.yahoo.com/cybersecurity/articles/scoop-top-ai-companies-probing-223553422.htmlAustralian inquiry asks Altman and Amodei to testify
https://thenextweb.com/news/australia-senate-inquiry-altman-amodei-openai-medicareMilitary personnel data exposed in breach, agency warns
https://www.militarytimes.com/news/pentagon-congress/2026/09/24/military-personnel-data-exposed-in-breach-agency-warns/Apple owes Taction $5.7B after losing haptic feedback IP trial
https://appleinsider.com/articles/26/09/26/apple-owes-taction-57b-after-losing-haptic-feedback-ip-trialInsurers claim AI is already increasing healthcare costs
https://techcrunch.com/2026/09/26/insurers-claim-ai-is-already-increasing-healthcare-costs/TikTok agrees to pay at least $100M in Alabama settlement
https://techcrunch.com/2026/09/26/tiktok-agrees-to-pay-at-least-100m-in-alabama-settlement/Anthropic now bills for refused Claude API requests, in three of five categories
https://mixed-news.com/en/anthropic-bills-refused-claude-api-requests-three-categories/Claude Opus 5.5 uses 95% fewer em dashes, but its answers are getting longer
https://www.bleepingcomputer.com/news/artificial-intelligence/claude-opus-55-uses-95-percent-fewer-em-dashes-but-its-answers-are-getting-longer/Sign-off