OpenAI Rolls Out Advanced Account Security for High-Risk Users

OpenAI just dropped a major security upgrade for ChatGPT and Codex accounts, targeting journalists, activists, and anyone who might be targeted by digital attackers. This opt-in setting, called Advanced Account Security, packs several beefed-up protections into one place. It forces passkeys or physical security keys for login—no more password-based sign-ins. Account recovery gets stricter too: no emails or SMS, just backup passkeys and recovery keys. Sessions are shorter to limit exposure if your device gets compromised, and you’ll get alerts on new logins. The catch is that OpenAI Support won’t help with recovery if you lock yourself out, so you must keep those recovery keys safe. Why it matters: As AI accounts become treasure troves of personal and professional secrets, this move signals that the industry is finally treating account security with the seriousness it deserves—even if it means shifting more responsibility onto users.