Microsoft today announced the launch of Face Check, a new facial recognition capability for its Entra Verified ID identity verification service. Face Check is designed to enable businesses and organizations to perform highly secure identity checks, while also protecting user privacy.
The new Face Check feature provides real-time facial matching between a user's selfie and the photo on their identity document, such as a passport or driver's license. Powered by Azure AI services, Face Check can match the live image against the individual's verified credentials to ensure they are who they claim to be before granting access to sensitive information or services.
Crucially, Face Check only shares the facial matching results and confidence score with the verifying organization. The user's actual selfie or biometric data is not stored or transmitted, helping preserve privacy while still allowing identity confirmation.
Microsoft says that Face Check can help enterprises reduce identity impersonation and fraud when providing employee access to privileged systems. The feature is currently being used by customers like BEMO, a leader in cybersecurity help desk services, to securely verify employees before provisioning admin access.
Face Check is distinguished from more controversial facial recognition systems used for surveillance or investigations without consent. As a voluntary identity verification method, Microsoft stresses that Face Check aligns with its responsible AI principles around privacy and transparency.
The Face Check capability for Entra Verified ID is now in public preview. Microsoft plans additional verification partnerships and attributes in the future as it seeks to establish Entra as an open-standards based identity validation platform.
With AI-enabled identity fraud on the rise, Face Check aims to balance enhanced security with ethical considerations around biometrics for enterprises navigating digital transformation. Its privacy-focused approach helps mitigate risks as facial analysis becomes an increasingly common authentication mechanism.