OpenAI said on August 19 that it has begun testing Private Safety Processing with early customers using its advanced AI models.

The system is designed to detect risky behavior across multiple interactions while keeping customer content inaccessible to OpenAI.

The system is meant to spot patterns that may not be visible in a single prompt-response exchange as AI agents take on more multi-step tasks.

OpenAI also said some customers using its application programming interface, or API, currently have “zero data retention,” meaning the company does not retain prompts or model responses after processing a request.

Early testing includes Microsoft and Databricks, a US-based data and AI software company.

OpenAI said it expects to roll out the system and publish a technical white paper in September.

🔗 Source: Bloomberg

Recent OpenAI developments