OpenAI Launches Private Safety Processing to Protect Customer Privacy
· Technology · TechCrunch, Indian Express
OpenAI announced a new service called Private Safety Processing that can flag misuse of its models without retaining any customer data. The system builds on OpenAI’s existing Zero Data Retention framework, extending the monitoring scope across multiple sessions rather than just individual interactions. The move comes after Anthropic introduced a data‑retention policy that keeps all conversations for 30 days for its covered Mythos‑class models, a policy that some enterprise customers have criticised. Private Safety Processing uses automated agents to analyse inputs and outputs across conversations, enabling detection of malicious activity such as malware engineering that spreads across separate requests. The feature is positioned as a competitive edge over Anthropic’s approach, signalling OpenAI’s focus on privacy‑friendly safety mechanisms.
Why it matters
Enterprises that process sensitive data could adopt OpenAI models while keeping their conversations off‑premise. The new system may alter how AI firms balance safety monitoring with privacy and could influence regulatory scrutiny in data‑protected jurisdictions.
Read the original report — TechCrunch
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.