Advanced Account Security Introduced
Key point
OpenAI has introduced Advanced Account Security for ChatGPT and Codex.
Details
OpenAI has added Advanced Account Security to ChatGPT's web security settings. It is an opt-in setting for users at high risk of digital attacks and those who want the strongest protection, and once enrolled, it also applies to Codex accessed with the same login.
This feature strengthens login, recovery, session management, and training exclusion all in one place.
- Login protection: Turns off password login and allows only passkeys or physical security keys.
- Recovery protection: Blocks email/SMS recovery and allows recovery only via backup passkeys, security keys, or recovery keys. After enrollment, account recovery support from OpenAI Support is also unavailable.
- Session management: Shortens session duration and provides login alerts and active session management features.
- Automatic training exclusion: Conversations from accounts with this setting turned on are automatically excluded from model training.
In partnership with Yubico, preferred pricing is also offered for YubiKey C Nano and YubiKey C NFC bundles. These bundles are also offered to eligible users from the security settings on the web, with the C Nano designed for everyday use left plugged into a laptop, and the C NFC designed for backup and dual use with laptops and mobile devices. Other FIDO-compatible security keys or software passkeys can also be used.
Additionally, individuals covered under Trusted Access for Cyber must turn on this setting starting June 1, 2026 in order to use the strongest cyber models. Organizations can instead demonstrate that they have phishing-resistant authentication in place within their SSO flow.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.