OpenAI previews Private Safety Processing, continuing to support cutting-edge ZDR models.
OpenAI has confirmed that it will continue to provide Zero Data Retention for cutting-edge models and preview Private Safety Processing to identify cross-interaction risks without giving OpenAI personnel access to the underlying content.
OpenAI confirmed that it will continue to offer the cutting-edge Zero Data Retention model to eligible API customers and preview Private Safety Processing.
The new solution aims to identify security risks from relevant interactions while preventing OpenAI personnel from accessing underlying content. OpenAI stated that it has already begun testing with early customers and plans to roll it out gradually in September, along with the release of a technical white paper; specific eligibility and endpoint applicability are subject to official data policies.
source:OpenAI official statement。