OpenAI previewed Private Safety Processing, an automated safeguard system designed to detect misuse across multi-turn interactions while maintaining Zero Data Retention protections.

Key Details

• Identifies safety risks across related interactions without exposing underlying prompt or response content to OpenAI staff.
• Supports customer-controlled infrastructure and OpenAI-hosted storage encrypted with customer-managed keys.
• Currently in early testing with enterprise customers, with full rollout and a white paper scheduled for September.

In This Story