OpenAI has introduced a major security upgrade for ChatGPT users with an opt-in ‘Advanced Account Security’ system. Designed to protect against phishing and account breaches, the feature targets high-risk users such as journalists, researchers, and public figures and remains open to anyone seeking stronger safeguards.
The system replaces traditional passwords with passkeys and hardware security keys. It offers greater protection against phishing attacks. “We are excited to announce our latest innovation for protecting users from malicious cyberattacks,” said OpenAI. It shows that device-based authentication now takes precedence over passwords.
Convenience is being sacrificed for reliability. Users who opt in will lose email and SMS recovery options, and will have to depend on backup keys or passkeys. This shift is expected to polarize users, balancing stronger protection against reduced flexibility.
In addition to identity verification, the platform offers reduced session duration, notifications about any anomalies, and full transparency on active sessions. Chats from accounts that have enabled ‘Advanced Account Security’ will not be utilized in the model's training process.
The company has partnered with Yubico to provide cheaper YubiKeys to increase adoption. As part of updated global security standards, certain customers are required to use the Trusted Access for Cyber program starting June 1, 2026.
OpenAI’s latest move represents an emerging trend within the industry that aims to reduce reliance on passwords in sensitive settings. This update ensures greater security in account management, and makes individuals more responsible when it comes to safekeeping of access credentials.
Also Read: OpenAI's ChatGPT 5.5 Enters Final Safety Review With April Release Window