OpenAI launches Privacy Filter, an open source, on-device data sanitization model that removes personal information from enterprise datasets

OpenAI released Privacy Filter, a 1.5B open-source model on Hugging Face for on-device PII detection and redaction.

It can run locally or in a browser to scrub sensitive data before enterprise text is sent to cloud systems.

The model helps organizations reduce exposure risks while supporting privacy and regulatory compliance like GDPR and HIPAA.

OpenAI says this approach preserves AI usefulness while keeping more sensitive information under customer control.