Google adds video selfies as an account recovery option

Google has introduced selfie video sign-in as an additional way for users to regain access to their accounts, the company said in a Thursday blog post dated July 23. The feature gives users another authentication route when they are locked out or cannot use their regular phone or computer.

The system asks users to record a short video through their device camera, using guided head movements so Google can capture their face from several angles, according to the company. If the user later has trouble signing in, Google said it can compare a new selfie video with the stored version to help confirm the account holder’s identity.

The company described the tool as an account access feature rather than a replacement for its existing security controls. Google said it will continue to apply its standard protections for suspicious sign-in attempts while adding checks designed to reduce impersonation risk.

How the feature works

Video-based verification relies on two elements described by Google: comparison with the previously saved selfie video and a liveness check. The comparison tests whether the person attempting to sign in matches the enrolled account holder. The liveness step requires the user to make simple movements, which Google said is intended to help distinguish a live person from an image or manipulated recording.

Google said the design addresses attempts to use deepfake photos and videos. The company did not provide technical detail on the models or thresholds used for those checks, but said the process includes multiple layers of protection against impersonation.

The feature also introduces a new category of stored identity data inside Google accounts. Google said selfie videos are recorded and stored only with user consent, can be deleted at any time in the user’s Google Account, and are used only to help with sign-in unless the user chooses to share them for other purposes. The company said the videos are encrypted at rest, meaning they are protected while stored and not actively being used.

AI raises pressure on identity controls

The launch comes as companies face more sophisticated identity fraud enabled by artificial intelligence. PYMNTS recently reported that scammers can use stolen identities, deepfake technology and remotely controlled devices to create “synthetic insiders” who appear to pass hiring, background check and access procedures.

That report said such actors may log in through approved accounts, creating access that looks legitimate because the credentials and permissions are valid, even if the person operating them is not who an employer believes it hired.

PYMNTS also reported in May on identity and authentication risks in financial services through an interview with Henry Patishman, executive vice president for identity verification solutions at Regula. Patishman said financial services authentication systems were built around the assumption that one person initiates an action, one system verifies that person, and the check occurs at a single moment. He said that assumption has weakened as fraud methods have changed.

Google’s new sign-in option reflects a broader shift in consumer authentication toward tools that combine possession of an account, biometric comparison and liveness checks. The company framed selfie video sign-in as an optional recovery method, with deletion and consent controls for users who choose to enroll.

This story draws on original reporting from PYMNTS.