Google introduces selfie video login for account recovery

New authentication feature allows users to verify their identity with a short selfie video when signing into their Google accounts.

Google has launched a new account recovery feature that allows users to verify their identity with a selfie video, offering an alternative way to regain access if they forget their password.

The new authentication method requires users to first enroll by recording a short selfie video. During the setup process, users are guided through a series of simple head movements while looking at their device’s camera, allowing Google to capture multiple facial angles for verification.

Once enrolled, users can sign in to their Google account by recording another selfie video whenever identity verification is required, such as during account recovery. Google compares the new recording with the original enrollment video to confirm the user’s identity before granting access.

According to Google, users remain in control of their biometric data and can delete the stored selfie video from their account settings at any time.

The company said the recorded video is used solely for account sign-in unless users explicitly choose to share it for additional purposes.

Google also emphasized that the stored selfie videos are encrypted while at rest, meaning the data remains securely protected even when it is not actively being used.

To guard against fraud, Google said the feature incorporates multiple layers of security designed to detect and prevent impersonation attempts using fake photos, deepfakes, or manipulated videos.

The selfie video authentication system is the latest addition to Google’s account security tools, complementing existing options such as passkeys, security keys, two-factor authentication, and biometric verification available on supported devices.

The move reflects Google’s broader efforts to strengthen account security while making the sign-in process more convenient for users who may have difficulty accessing their accounts through traditional password-based methods.