Digital rights activists are demanding a ban on government use of facial recognition technology, after it emerged that US authorities mined driver’s-license picture databases to conduct undisclosed surveillance and immigration enforcement.

Advocacy group Fight for the Future argues facial recognition is “biased, invasive, and violates basic rights [...] to spy on the American public.”

Studies have found the artificial intelligence—often trained on Caucasian features—misidentifies darker-skinned individuals. In cities like Detroit or London, citizens are alarmed at the technology’s inaccuracy.

Two US cities have already barred local agencies from deploying facial recognition.

However, some experts believe a complete federal ban remains improbable, particularly given the involvement of heavyweights such as Microsoft and Amazon, seeking to get ahead of the curve by calling for regulation rather than prohibition.

Beyond infringing privacy and jeopardising presumption of innocence, critics also stress facial recognition’s cybersecurity risk.

They cite a May hack on Customs and Border Protection databases which captured 100,000 Americans’ biometric information. CBP had been collecting visa and passport photos to develop its facial recognition programme for airports.

But advocates say the technology—once improved—could help swiftly localise criminals, definitively inculpate or exculpate suspects, and prevent data breaches.

How should we balance privacy with security?