OpenAI chases Anthropic's biz customers with zero data retention pledge
With Private Safety Processing, AI biz promises discreet, automated prompt surveillance
OpenAI has introduced a new feature called Private Safety Processing, which allows customers to retain no data while still ensuring AI model safety. This move could prove to be a significant advantage over rival Anthropic, which has yet to implement a similar system. OpenAI's latest offering, Private Safety Processing, enables automatic scanning of customer model interactions for potential safety risks without violating Zero Data Retention commitments.
This feature comes as Anthropic has disclosed that it retains data, albeit in limited quantities, for its covered models as part of its safety measures. According to OpenAI, the free-spending AI company's Private Safety Processing is designed to automatically scan customer model interactions for safety risks without compromising Zero Data Retention commitments.
Anthropic, on the other hand, has acknowledged that its implementation of Zero Data Retention includes a non-zero amount of data retention for covered models. OpenAI's Private Safety Processing is set to be explained in more detail next month, with the company emphasizing that whatever data is stored is inaccessible to its personnel.
The company's privacy commitment extends to the expanded interest in related interactions, suggesting that OpenAI's safety assessment will examine not only prompt input and output but also tool use and network data signals. However, privacy expert Matthew Green from Johns Hopkins University has cautioned that private inference alone is not sufficient to ensure data protection, as AI agents often rely on sensitive data when acting on behalf of users.
Written by urgent.news from The Register Science's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
