OpenAI Introduces Advanced Account Security for ChatGPT Users

By: Harsh Vardhan

|

Updated on: 01-May-2026 09:01 PM

1,289 views

Follow Us:

1,289 views

OpenAI has launched a new feature called Advanced Account Security for ChatGPT users. This opt-in setting aims to protect individuals who handle sensitive information, such as journalists, researchers, and public figures. The feature is also available to any user seeking higher security for their accounts.

Key Highlights

• OpenAI launches Advanced Account Security for ChatGPT users handling sensitive information.
• Feature replaces passwords with passkeys or physical security keys for stronger protection.
• Account recovery now requires backup keys or passkeys instead of email or SMS.
• Conversations with this setting enabled are excluded from model training for added privacy.
• Trusted Access for Cyber users must enable this feature starting June 1, 2026.

Key Features of Advanced Account Security

The new security system replaces traditional password-based logins with more secure methods. Users can now use passkeys or physical security keys to access their accounts. This change makes accounts more resistant to phishing attacks, according to OpenAI.

Account recovery methods have also changed. Email and SMS-based recovery options are no longer available for users who enable this feature. Instead, recovery now requires backup keys or passkeys. This approach strengthens security but means users cannot rely on standard support channels for account recovery.

Login session durations have been shortened to reduce the risk of unauthorized access. Users will receive real-time alerts if suspicious activity is detected on their accounts. They can also review and manage all active sessions across devices, giving them more control over account access.

Privacy and Adoption Initiatives

Conversations from accounts with Advanced Account Security enabled will not be used for model training. This offers an extra layer of privacy for users who opt in to the feature.

OpenAI has partnered with Yubico to encourage the use of phishing-resistant authentication methods. The partnership will provide discounted hardware security keys, including YubiKeys, to users who want to adopt these stronger protections.

Starting June 1, 2026, users enrolled in OpenAI's Trusted Access for Cyber programme must enable Advanced Account Security. This requirement aims to further protect users at higher risk of digital attacks.

Background and Availability

OpenAI announced the new feature in an official blog post. The company emphasized that the setting is designed for those at increased risk of digital attacks, but it remains available to any user who wants stronger protections. The feature is now accessible to all ChatGPT users who choose to opt in.