FBI Facial Recognition Software Identifies The Wrong Person Up To 20% Of The Time

In 2012, the Electronic Privacy Information Center (EPIC) filed a Freedom of Information Act request to the FBI. EPIC filed the request in hopes of attaining documents related to the FBI’s Next Generation Identification (NGI) biometric system.

Long story short, NGI is a way of identifying people. It’s a collection of fingerprints, iris scans, DNA profiles, images, etc. designed to help intelligence and law enforcement agencies identify whoever they need to.

As part of their Freedom of Information Act (FOIA) request, EPIC obtained documents detailing the accuracy of the facial recognition software used in NGI. As it turns out, the accuracy standard for the software is pretty low:

“NGI shall return the correct candidate a minimum of 85% of the time, when it exists in the searched repository, as a result of facial recognition search in support of photo investigation services.

"NGI shall return an incorrect candidate a maximum of 20% of the time, as a result of facial recognition search in support of photo investigation services.”

Although a 1-in-5 chance of incorrect identification seems pretty high, you’re not likely to be called down to the police station anytime soon for a crime you know nothing about. The facial recognition system is used in tandem with the other identification technologies, most of which have much higher accuracy rates. Iris scanning, for example, returns the correct candidate a minimum of 98% of the time. Fingerprint scans are claimed to return the correct candidate 99% of the time.

You’re opinion about the NGI program as whole, however, is another issue. 

Source: EPIC, Reason


Popular Video