OpenAI has introduced a new privacy-centric approach to monitor AI misuse, aiming to differentiate itself from rival Anthropic. The company is previewing a service called Private Safety Processing for select customers. This automated system detects potential abuse without retaining any customer data, directly countering Anthropic’s recently announced data-retention policy. The service is designed to identify malicious use across multiple sessions, a spokesperson told TechCrunch. The system uses agents to analyze interactions and send narrowly defined signals to OpenAI when suspicious activity is detected. OpenAI can then decide whether to engage with the customer for further context or take enforcement action.
Private Safety Processing expands upon OpenAI’s existing Zero Data Retention policy. Unlike Anthropic, which retains data for covered models like Fable, OpenAI’s new technology allows for long-horizon safety monitoring. This means the system can assess inputs and outputs from multiple conversations, rather than just single sessions. The approach enables the company to detect patterns of misuse that might otherwise go unnoticed. If triggered, the system sends a signal to OpenAI, which then determines whether to involve the customer or take further action. The company emphasized that customers retain discretion over whether to share data with OpenAI.
Anthropic’s data-retention policy, which allows for 30 days of user data storage for covered models, has raised concerns among enterprise clients. The company states that human review of customer data can occur through a controlled access path involving a small set of approved reviewers. Every review session is recorded in a tamper-proof log that cannot be altered. This contrasts with OpenAI’s approach, which prioritizes privacy by not retaining any data. The competitive tension between OpenAI and Anthropic continues as both companies seek to gain an edge in the AI market. OpenAI’s Q2 growth was slower than Anthropic’s, with the latter’s annualized revenue run rate now reportedly $65 billion.
Source: techcrunch
