OpenAI announced the launch of a new feature called “Advanced Account Security” to raise the level of protection for ChatGPT accounts. It allows users to secure their accounts using physical security keys instead of relying on traditional passwords.
The feature, which can be activated optionally, provides a complete package of protection measures without the need to navigate between multiple settings. When activated, password login is completely eliminated, replaced by what are known as “passkeys” or physical security keys.
In addition, the duration of active sessions is reduced, which limits the exploitation of any suspicious login, with immediate alerts sent when an attempt is made to log into the account.
The company excludes conversations associated with accounts that have enabled the feature from being used to train models automatically, without the need to manually adjust this option.
The feature imposes fundamental changes in account recovery mechanisms, as traditional recovery methods via email or text messages are disabled. If the security keys are lost, the user will not be able to regain access to them even through company support.
“OpenAI” explains that this trend aims to reduce the risks of hacking associated with hacking email or phone numbers, which are among the most common methods of seizing accounts.
In a related context, the company entered into a partnership with Yubico, a company specializing in cross-device authentication solutions, to provide low-cost key packages produced by it. The packages include two keys; One is small and is almost always used with computers, and another supports NFC technology for use with mobile phones.
The feature may not seem essential to all users currently, but it could be useful to users who rely on ChatGPT for sensitive tasks such as legal research, medical inquiries, and business, making their accounts increasingly valuable targets.
