OpenAI announcement · Published 2026-08-19
OpenAI previews safety processing across interactions that stays compatible with Zero Data Retention
OpenAI describes zero data retention for eligible API requests and private-processing research across interactions.
Retain ZDR while examining cross-interaction risks
OpenAI describes Private Safety Processing as a way to identify risk patterns across related interactions. The company says eligible API customers retain the ZDR commitment that prompts and responses are not kept after processing, with underlying content unavailable to OpenAI personnel.
The testing stage in the source record
The August 19, 2026 source describes early customer testing, a planned September rollout and the statutory image-review exception. This records the scheme and plans at that time.
Sources and notes
sources中文
Sources and notes
- Read the originalOpenAI announcement
Retained summary and context
Verified key facts: ZDR keeps its promise that prompts and responses are not retained after processing; the new Private Safety Processing only flags risk across related interactions without exposing underlying content to OpenAI staff; it is testing with early customers with rollout planned from September; the legal human-review exception for CSAM imagery remains.
It is still being tested with early customers, rollout is planned for September, and the legal manual-review exception for flagged CSAM images remains.
Sources and evidence · 1 link
Each link has a purpose and scope. Link counts do not establish independent confirmation; direct support applies only to the statements specified below.
Source timeline
- · OpenAI published the Private Safety Processing preview and its ZDR boundaries.
- · Rollout and a technical white paper are planned, not yet delivered.
Collection record
First discovered:
Collection and source publication have distinct dates.