Observed facts: OpenAI’s feature removes selected conversations from regular history, pinned chats, and conversation search, and places them in a separate project. Access requires a four-digit PIN, with device authentication such as Face ID available on iOS and Android; resetting the PIN deletes chats and uploaded files without recovery.
Mechanism: The control limits casual visibility on a device without changing service-side retention, training settings, or legal obligations. It is an interface-level barrier rather than an encryption or data-governance change.
Why now: Chat histories increasingly accompany users into shared-device and screen-sharing contexts, making accidental exposure a practical interface-level concern.
First implication: Treat the feature as a privacy convenience, not as end-to-end encryption or confidentiality from OpenAI. Users should review data settings separately; the lock does not change those controls.
Evidence boundary: The supplied Engadget report describes the feature and cites OpenAI support information; it does not independently test implementation, rollout timing, or enforcement.
First watch condition: Watch whether OpenAI documents changes to retention, training controls, or recovery behavior for locked chats. Those would indicate a shift beyond interface-level privacy.
This is a distinct analytical record because no prior New Runtime record on chat locking or this specific access-control boundary was supplied.