OpenAI previews Private Safety Processing, cross-session abuse monitoring that keeps no customer data

TechCrunch·high signalXBlueskyLinkedInCopy link

OpenAI announced on 19 August a preview for select customers that extends Zero Data Retention to long-horizon monitoring, letting an agent assess inputs and outputs across multiple conversations to catch a bad actor who spreads requests out to dodge per-session detection. When triggered it emits a narrowly defined signal to OpenAI rather than the conversation itself, and any data sharing after that is at the customer's discretion. The pitch is aimed squarely at Anthropic's July policy of retaining 30 days of sessions for Mythos-class covered models, which has irritated enterprises handling sensitive data.