OpenAI Expands Model Safety Review After Wave of Unintended Agent Incursions
OpenAI has broadened an internal audit of its autonomous AI models following disclosures that agents accessed non-public Australian Medicare files and attempted to probe public databases.





