Daily AI news, sourcedUpdates / Policy & safety
Updates Policy & safety · 115 items · newest issue first · page 6 of 6
AreaAllModelsAgents & codingApps & productsChips & infraResearchPricing & limitsBusiness & fundingPolicy & safetyOther
- OpenAI lays out safety-case approach for frontier AI training
- 'It's not just the sandbox': perspective from an internal security person at OpenAI
- OpenAI apologises to Australia and pledges stronger cyber safeguards
- Artificial Analysis launches Cyber Index for AI agent security
- OpenAI brings frontier RL training into its safety-case review process
- Trump and Mike Johnson to host six tech CEOs at White House AI lunch
- OpenAI: research agents sent data to third parties, incl. images in 53 cases
- OpenAI and Anthropic are investigating tens of thousands of model incidents, per Axios
- China weighs letting ByteDance and Alibaba buy Nvidia's RTX PRO 5500, per The Information
- NVIDIA unveils Open Agent Safety Platform, pairing open-source OpenShell with Sentry on BlueField-4OfficialNVIDIA Developer blogSep 28Policy & safety
- Anthropic will skip Thursday's Australian Senate hearing on rogue AI agents, per The Sydney Morning HeraldReportedThe Sydney Morning HeraldSep 28Policy & safety
- China now requires Beijing's approval for trips abroad by spouses and children of some top AI and chip executives, per Bloomberg
- Contractors rating Microsoft Copilot image edits could see users' prompts and uploaded photos, per 404 MediaReported404 MediaSep 28Policy & safety
- Florida asks court to bar OpenAI from building new models without outside oversight, per ReutersReportedReutersSep 28Policy & safety
- UK AISI: GPT-6 Astra ran out-of-scope supply-chain attacks more often than earlier OpenAI models in simulated testsOfficialUK AI Security InstituteSep 28Policy & safety