ChatGPT’s security is becoming increasingly important as it accumulates sensitive information that you provide. That’s why we created Advanced Account Security—a voluntary option to protect your account.
The more time you spend on an app, the more likely it is that it will accumulate personal data and sensitive information that you don’t want falling into the wrong hands. Your email client, your photo gallery, and WhatsApp are three prime examples. But there are more. Like ChatGPT. For many, it’s a personal assistant, a work tool, or much more. With all of them in mind, OpenAI has created a ChatGPT security feature called Advanced Account Security .
As we’ll see later, OpenAI already offers security features in ChatGPT to prevent your account from falling into the wrong hands or your data from ending up where it shouldn’t. These include, for example, passkeys , multi -factor authentication (MFA) or two-step verification (2FA), and the ability to log out of your session on all devices where you use this AI assistant. But nothing is ever truly enough when it comes to cybersecurity .
So, optionally, if you believe the information stored by ChatGPT or Codex is sensitive or too important to allow unauthorized access, you can request that OpenAI activate the Advanced Account Security option in your user account . Let’s see exactly what this does and what you need to request this security measure.
What is Advanced Account Security?
Advanced Account Security is “a new voluntary activation option for ChatGPT accounts, designed for people at higher risk of cyberattacks , as well as those who want the most effective protections available for their account.”
OpenAI then explains the reason for this feature. “Over time, a ChatGPT account can contain sensitive personal and professional context and become central to connected tools and workflows.” It mentions professional profiles such as “journalists, elected officials, political dissidents, researchers , ” and, in general, “those who are particularly security-conscious.”
Advanced account security consists of several controls or elements focused on various important aspects of your user account: logging in, account recovery, and compromised sessions. When you activate this security measure, OpenAI does the following in ChatGPT and Codex :
- Deactivate login codes via email and SMS.
- Disable password login .
- Disable account recovery via email.
- Deactivate model training .
- Activate email alerts for login attempts.
- Activate shorter active sessions .
With Advanced Account Security enabled in ChatGPT, when logging into your user account, you’ll need to use digital access keys or physical security keys . You won’t have to remember your password. And your account will be more resistant to phishing. To recover your account, you’ll need to use the same key method, instead of SMS or email. Third, sessions will be shorter. You’ll need to authenticate more frequently . And with each authentication, you’ll receive a notification to prevent unauthorized access. Finally, those who work with particularly sensitive information can opt out of having those conversations used for model training.
How to activate Advanced Security
By default, Advanced Account Security is disabled. Not everyone needs this level of security. But if you’re one of those who do, simply request it. You’ll find this option when you log in to ChatGPT or Codex . Go to Settings , which you can access by clicking or tapping on your user profile. In Settings, go to Security. There you’ll find Advanced Account Security . Click Enroll , read the warnings beforehand, click Enroll again , and follow the instructions.
To activate this feature, you’ll need to do so through the ChatGPT web version . While the apps mention the feature and indicate whether it’s enabled, they don’t provide any further information. Once activated, it will work across all your open apps and sessions . Furthermore, you can use Advanced Account Security on any ChatGPT plan, free or paid, except for the Enterprise plan . It’s also available in most countries.
Before requesting that OpenAI activate ChatGPT’s extreme security, the official help documentation recommends configuring at least two login methods. These are access keys or passkeys , physical security keys, or a combination. Activating one of these methods will generate recovery keys . Keep these safe in case you need them.
Other ChatGPT security measures
You might find the Advanced Account Security feature too restrictive to safeguard the data and conversations you store in ChatGPT. If so, you still have other, less drastic options available. As we saw earlier, in the ChatGPT Security section , you can enable other methods to log in more securely than simply using a password.
- Multi-factor authentication (MFA) . Also known as multi-factor verification, multi-factor identification, two-step authentication, multi-factor login, two-factor authentication, etc. The idea is to introduce a second verification of your identity . Once activated, in addition to using your usual password, you will need a second system to confirm that it is you logging into ChatGPT. This can be an authenticator app like Google Authenticator or Authy , a notification in the ChatGPT app on your trusted device, or an SMS or WhatsApp message with a 6-digit code.
- Security keys and passkeys . The former are physical, like Yubico ‘s YubiKeys . The latter are digital. A passkey replaces a password with a more secure and convenient method, such as fingerprint or facial recognition . When you create the passkey, your device stores a private key, and the online service, in this case ChatGPT, receives a public one. This way, the password is never sent and cannot be easily stolen through phishing. Apple, Microsoft , and Google allow you to create and use this system, so you can use it on Windows, macOS, Android, or iOS.
- Close all sessions . Another ChatGPT security option that can be useful if you detect that someone has accessed your user account is to close all open sessions. This way, you can then change your password, activate other, more secure login methods, and ultimately prevent unauthorized access. You’ll find this option in ChatGPT by going to Settings > Security > Close all sessions .
