Observed facts: OpenAI says locked chats leave regular history and conversation search and require a four-digit account PIN, with supported device authentication on some mobile devices. Mechanism: The feature relocates selected conversations into a dedicated Locked chats area and gates opening them behind account authentication; it does not encrypt them as a separate vault. Why now: The rollout gives users a built-in way to keep sensitive conversations out of view on shared screens or devices, while making the distinction between concealment and data protection consequential.
The Help Center explicitly says locking does not change retention or the model-improvement setting. Resetting a forgotten PIN permanently deletes locked chats and content protected only by that PIN. First implication: Teams should treat this as a convenience control for reducing casual exposure, not as a substitute for workspace policy or a confidentiality boundary. The source also notes that other already-open devices can temporarily retain titles or loaded content, so locking may not immediately clear every surface.
Evidence boundary: The evidence is OpenAI’s Help Center description; it establishes documented behavior and rollout status, not independent security testing or universal availability at a given moment. First watch condition: Watch whether OpenAI clarifies the sync/refresh behavior across devices and offers stronger administrative controls for managed workspaces. This begins a distinct analytical record because no prior New Runtime record is supplied for this specific access-control mechanism.