If your business uses frontier models through an API, the latest privacy update is not just policy wording. It changes what you should assume about request retention and how to think about safety processing for sensitive workloads.
What changed
OpenAI reaffirms Zero Data Retention for eligible API customers. OpenAI also previews Private Safety Processing for advanced AI safety checks while aiming to avoid compromising data privacy for those safety steps.
Why it matters for business teams
Zero Data Retention reduces the risk that your prompts and outputs are kept for later use. Private Safety Processing is designed to add safety coverage while keeping the privacy promise aligned with the Zero Data Retention approach for eligible customers.
What to do next
Check whether your API usage qualifies for Zero Data Retention. Then review your internal data handling assumptions for safety related processing, and update your model use documentation so teams know what to expect for retention and privacy boundaries.