OpenAI Unveils ‘Private Safety Processing’ to Protect Enterprise Data

OpenAI has previewed Private Safety Processing, a new system aimed at strengthening AI safety while preserving privacy for enterprise and API customers. The technology is designed to identify patterns associated with potentially harmful or abusive AI activity across interactions without giving OpenAI employees access to the underlying customer content. The system builds on OpenAI’s existing Zero Data Retention (ZDR) framework, which is available to eligible API customers.

Under the new approach, customer data can remain within customer-controlled infrastructure, while OpenAI is also working on an option to store encrypted interaction data on its infrastructure with encryption keys controlled by the customer. When an automated safety system detects suspicious activity, OpenAI says it will receive only a narrowly defined signal about the type of activity rather than the actual customer content. Customers can investigate alerts within their own systems and can voluntarily share relevant information with OpenAI if they want to appeal a decision or assist with an investigation. The system is currently being tested with selected customers and is expected to become more widely available in September, alongside a technical white paper.