Google brings liveness detection to anti-bot verification

Google is testing a new reCAPTCHA method that asks users to perform simple hand gestures on camera. The service is part of Google Cloud Fraud Defense.
The biometric verification system analyzes short video clips of a user’s hand and extracts 21 knuckle‑point coordinates to determine human liveness. The approach reflects trends toward using anatomical or behavioral cues to distinguish humans from bots.
Google says the footage is processed in real time, never linked to an identity and deleted as soon as the verification ends. The company stresses that no images or videos are stored beyond the verification moment, and the extracted landmark data is not reused for training or other purposes. Audio is never recorded.
All information collected falls under the Google Privacy Policy, and camera access requires explicit user consent. Permissions can be revoked at any time through browser settings, and Google says it does not share gesture‑related data or permissions with third parties.
Google doesn’t describe the feature as biometric authentication, but the extraction of anatomical landmarks places it in a regulatory grey zone at a time when UK and EU authorities are scrutinizing even ephemeral biometric processing. Google is emphasizing data minimization with no identity linkage, retention or secondary use.
The gesture check is a privacy‑preserving evolution of CAPTCHA rather than a wholly new biometric system. Whether regulators and privacy advocates accept that distinction will inform how widely the feature is deployed. Some users online have expressed concern about the use of a camera, saying it’s more intrusive than traditional CAPTCHAs.
Lokesh Sparrow on X argued the approach raises “serious privacy concerns” as it requires camera access for routine verification. “It also raises real accessibility issues for people with hand disabilities or those in poor lighting conditions, making basic web access more difficult for many users,” he added.
Google says it is developing additional accessible alternatives and that users who cannot perform gestures will still be offered visual or audio challenges.
Article Topics
biometric liveness detection | biometric verification | gesture recognition | Google | Google Cloud | hand biometrics | smartphones







Comments