AI more reliable than people at identifying biometric spoofs

Friday 11 February 2022 14:49 CET | News

According to a research report conducted by IDR&D, computers are more prone than people to determine the legitimacy of a person’s photo accurately and quickly versus a presentation attack.

As fraudsters attempt to imitate real-life customers during an online purchase process, including creating a new bank account or logging into an existing account, AI machines faster detect whether a photo taken in real time belongs to a real person.

The study tested both humans and machines by presenting them with the most common fraud-attempt technologies, including printed videos, 2D and 3D masks, and digital photos. As per the study, computers were 10 times quicker than their human peers to recognize a photo of an actual person or a spoof, while scoring 0% error rates across the 175,000 images provided and all types of attacks.

Passive facial liveness detection was also performed more accurately by AI by comparison to humans. The latter misclassified 18% of genuine faces as spoofs, as opposed to only 1% misclassifications from AI machines. 

More: Link

Free Headlines in your E-mail

Every day we send out a free e-mail with the most important headlines of the last 24 hours.

Subscribe now

Keywords: identity theft, identity verification, identity fraud, report, artificial intelligence, machine learning, digital identity
Categories: Fraud & Financial Crime
Countries: World
This article is part of category

Fraud & Financial Crime

Industry Events