OpenAI Launches Private Safety Processing to Detect Risks Under Zero Data Retention

OpenAI has introduced a preview of "Private Safety Processing," an automated feature that detects multi-interaction safety risks for eligible frontier model API customers. This system analyzes patterns across conversations and accounts without retaining raw customer prompts or responses or exposing them to OpenAI employees. This feature resolves a critical conflict between enterprise data privacy and AI safety monitoring, allowing organizations to maintain strict Zero Data Retention policies while still preventing abuse. It represents a significant step forward in secure enterprise AI deployment, especially as models handle longer and more complex tasks. When a risk is flagged, OpenAI only receives a limited safety signal describing the activity type, while the underlying content remains encrypted with customer-controlled keys. The feature is currently being tested with early customers, with a wider rollout and a technical whitepaper planned for September.