OpenAI builds safety system that catches misuse without storing customer data
decoder ·
OpenAI has developed a safety system called "Private Safety Processing" that allows enterprise customers to use its most advanced AI models without storing their data. The system detects abuse patterns across multiple related interactions while maintaining zero data retention, meaning no data is kept after processing. OpenAI only receives a narrow safety signal indicating the type and severity of activity, without seeing actual inputs or outputs. Customer data remains on the customer's own infrastructure or is stored encrypted, with the customer retaining the keys. Aleah Houze, Head of Product Policy at OpenAI, noted that risks often become apparent only over multiple conversations. This contrasts with competitor Anthropic, which requires 30 days of data retention for its most powerful models. A technical white paper is expected to be released in September.
decoder