OpenAI offers Zero Data Retention for frontier models, previews Private Safety Processing
OpenAI announced (Aug 19) Zero Data Retention for eligible frontier-model API deployments: prompts and responses are not retained after a request, personnel cannot review customer content, and enterprise data is not used for training unless customers opt in. The company also previewed Private Safety Processing—automated pattern detection across related interactions that returns narrow risk signals without exposing underlying content to OpenAI staff—with content staying on customer-controlled infrastructure or OpenAI storage encrypted under customer-held keys. Early-customer testing is underway ahead of a broader rollout and technical white paper planned for September; consumer ChatGPT Free/Plus/Go/Pro plans are unchanged. Distinct from prior ZDR docs and European data-residency posts.





