OpenAI tightens ChatGPT account security
- OpenAI launched Advanced Account Security for ChatGPT on April 30, bundling stronger login and recovery protections into one opt-in setting for consumer accounts. (openai.com) - Turning it on requires passkeys or compatible security keys, disables password and email/SMS code sign-ins, and adds recovery keys, login alerts, and shorter sessions. (help.openai.com) - The bigger point is that ChatGPT accounts now hold more sensitive work and personal context, while training-data controls remain a separate privacy setting. (openai.com)
ChatGPT accounts are starting to look less like throwaway logins and more like real digital vaults. People keep work drafts there, personal notes there, and increasingly conn(openai.com)I just added a new opt-in package called Advanced Account Security that hardens how people sign in, recover access, and manage active sessions. It went live on April 30, 2(help.openai.com)d. (openai.com) ### What actually changed? OpenAI bundled a set of stricter protections into one switch (openai.com)separate safeguards, Advanced Account Security turns on a higher-security mode meant for people at elevated risk of account takeover — journalists, executives, researchers, or really anyone who wants the strongest login setup available. (openai.com) ### Why is the login piece the big deal? Because the package cuts off weaker ways into the account. With Advanced Account Security enabled, ChatGPT uses passkeys or compatible hardwa(openai.com)codes, and email-based account recovery. Basically, OpenAI is pushing users away from channels that are easier to phish, intercept, or socially engineer. (help.openai.com) ### What happens if you get locked out? The catch is that stronger security means harsher recovery rules. OpenAI adds recovery keys, and users are supposed to save them before (openai.com)urity key and your recovery material, getting back in becomes much harder by design. That is the tradeoff — fewer weak recovery paths also means less wiggle room after a mistake. (help.openai.com) ### What about active sessions? OpenAI also tightened what happens after login. Advanced Account Security adds shorter active sessions, login notifications, and s(help.openai.com) more quickly. That matters because a lot of account compromises are not about guessing the password anymore — they are about stealing a live session and quietly riding it. (help.openai.com) ### Who can use it? Right now this is for eligible consumer ChatGPT accounts. OpenAI’s help docs say it is not available for ChatGPT Enterprise users, enterprise-managed accounts, or acco(help.openai.com)fessional” framing some coverage used, this is not an enterprise admin feature. It is a consumer-account hardening mode. (help.openai.com) ### Is this the same as turning off training? No — and that is an important distinction. Data-use controls live elsewhere. On personal ChatGPT plans, users can opt out of having new conversations used to improve mo(help.openai.com)Advanced Account Security. Business and enterprise privacy promises are handled under different product policies and controls. (help.openai.com) ### Why now? Because a ChatGPT account now carries more value than it did a year ago. OpenAI itself frames the issue around “deeply personal questions,” high-sta(help.openai.com)Add Codex access, shared context, and saved history, and the account starts to resemble a password manager crossed with a work notebook — not literally, but close enough that takeover risk becomes much more serious. (openai.com) ### So what is the bottom line? This is OpenAI admitting that AI accounts need security settings closer to banking or co(help.openai.com)rrower than some early summaries suggested — stronger sign-in and recovery, yes; a new bundled training-data exclusion feature, no. (openai.com)
