OpenAI tests new safety system for AI models to prevent misuse
OpenAI is testing a new safety system, Private Safety Processing, designed to detect AI misuse patterns without employee access to customer data. The system is currently in early testing and planned for a September rollout.
Artificial intelligence company OpenAI has begun testing a new safety system called Private Safety Processing. The system is designed to detect patterns of misuse across multiple AI interactions without granting OpenAI employees access to customer prompts or responses.
This new system is an extension of OpenAI's existing Zero Data Retention option, which allows eligible API customers to keep their prompts and model responses out of OpenAI's retained data. Private Safety Processing extends these protections to cover related interactions. OpenAI states that its automated systems can identify patterns that may indicate misuse, including activity that only becomes visible across multiple requests. Customers can continue to store content on their own infrastructure, or OpenAI is developing an option where content is stored on OpenAI infrastructure but encrypted with customer-controlled keys.
When the system detects a risk, OpenAI receives a limited signal about the type of activity involved, without access to the underlying content. Customers can review alerts through their own systems and choose to share relevant information with OpenAI when appealing an enforcement decision or supporting an investigation into verified abuse. An exception is made for images flagged as potential child sexual abuse material (CSAM), which will continue to be retained for manual review and reporting.
Competitor Anthropic employs a different approach, requiring a 30-day retention period for prompts and outputs submitted to certain models for safety monitoring, even for organizations using zero retention. OpenAI's new system aims to provide similar protections by detecting misuse patterns while keeping the underlying content outside of its employees' access.
OpenAI is currently testing the system with select early customers and plans to roll it out more broadly in September. The company has not yet specified whether Private Safety Processing will be available to Indian customers or if there will be any India-specific terms or restrictions upon launch.