The Advanced Account Security feature for ChatGPT accounts replaces traditional password logins with stronger authentication methods requiring physical security keys or passkeys. It improves account recovery by eliminating email- and SMS-based methods, instead using backup passkeys and recovery keys, though this limits support recovery options. Session lengths are shortened to reduce takeover risks, and users receive login alerts and session management capabilities. Additionally, conversations from enrolled users are excluded from AI model training. This security enhancement targets users at increased risk of targeted hacking and is opt-in.

The feature reduces the risk of account compromise by eliminating password-based authentication and less secure recovery methods. Shortened sessions and login alerts further mitigate account takeover risks. However, users must manage their recovery keys carefully as OpenAI support cannot assist with recovery once Advanced Account Security is enabled. The exclusion of user conversations from AI training protects user privacy and data confidentiality. There are no known exploits in the wild related to this feature.

Users at higher risk of targeted attacks are encouraged to opt in to Advanced Account Security to benefit from stronger authentication and recovery methods. Since this is an opt-in feature, no immediate action is required for general users. OpenAI manages this feature and provides a dedicated enrollment page. Users should securely store their recovery keys, as OpenAI support cannot recover accounts with this feature enabled. Patch status is not applicable as this is a new security feature rollout rather than a vulnerability patch.