About Face: When Biometric Recognition Gets it Wrong

Whether it’s a fingerprint or a face, biometric recognition systems are likely here to stay. 

Once solely within the purview of spy movies, biometric recognition is as common as the cell phone which often uses it to authorize use by its owner. We touch a screen, we look into it, and we gain entry into our communication devices, or perhaps our place of work.

But, let’s face it, technology isn’t foolproof. And when we are dealing with biometric facial recognition devices, we have to “face” the fact that sometimes there is an about face, and the technology gets it wrong — gets us wrong.

When did the technology first appear?‘The earliest pioneers of facial recognition were Woody Bledsoe, Helen Chan Wolf and Charles Bisson. In 1964 and 1965, Bledsoe, along with Wolf and Bisson began work using computers to recognize the human face.’[1]

The technology has made great progress since then of course, and as so often happens, when technology progresses, individual privacy rights regress.

Further, where there is new technology, law enforcement is close by to expand and exploit its use. Let’s look at the law enforcement use of biometric facial recognition by our fine federal friends.

‘Seven law enforcement agencies within the Departments of Justice (DOJ) and Homeland Security (DHS), such as the Federal Bureau of Investigation and U.S. Secret Service, reported using facial recognition technology to support criminal investigations. Three of the seven agencies reported owning facial recognition technology. All seven reported using systems owned by other entities, such as state and local entities and nongovernment service providers.’[2]

In the Government Accounting Office study, from which the previous quote comes, it was observed:

‘Civil rights and civil liberties advocates have cautioned that an overreliance on the technology in criminal investigations could lead to the arrest and prosecution of innocent people, or that its use at certain events (e.g., protests) could have a chilling effect on individuals' exercise of their First Amendment rights.’[3]

One can imagine the recording of peaceful protestors at a politically oriented rally, or the random but regular surveillance of a public area and the people passing by. But there’s no need to imagine it, because that’s a reality in today’s pretty-close-to Orwellian world.

Could biometric technology be a force in the promotion of police profiling of certain ethnicities? According to a PEW research study done in 2022, with respect to the respondents to survey questions:

‘66% say police would use this technology to monitor Black and Hispanic neighborhoods much more often than other neighborhoods. Americans are more divided on the effects facial recognition would have on false arrests. Some 53% of U.S. adults say police probably or definitely would make more false arrests if use of facial recognition technology was widespread among police.’[4]

So, the question arises, are there times when the technology goes astray and misidentifies someone? Are there times when biometric facial recognition either gets it wrong, or does it wrong with respect to individual expectations of privacy?

A study referenced within the PEW research study found that ‘A 2016 study out of Georgetown Law found that half of American adults’ faces were already in law enforcement’s facial recognition databases.’[5]

That was almost a decade ago, and it’s a safe — or unsafe — bet that while the technology has improved, the database has expanded to include by now most Americans. So much for the expectation of privacy.

It’s not so ‘expected’ anymore, because within that PEW study it was found that ‘Six-in-ten Americans say people should assume they are being monitored when they are in public spaces, while 39% say people should have a right to privacy when they are in public spaces.’[6]

What about biometric boo-boos, when the technology gets it wrong and misidentifies someone?

Can that, does that happen?

Yes.

In a 2024 Minnesota American Civil Liberties Union article on the subject, it was reported that ‘Studies show that facial recognition is least reliable for people of color, women, and non-binary individuals. And that can be life-threatening when the technology is in the hands of law enforcement.’[7]

Who is safest from potential biometric misidentification? Well, it may come as no surprise that, ‘A 2019 test by the federal government concluded the technology works best on middle-age white men. The accuracy rates weren’t impressive for people of color, women, children, and elderly individuals.’[8]

Interesting to note that it works best on the people who likely were the developers of the technology.

Why You Should Care

Even if you are a middle-aged white man, you no doubt are not in favor of you being a subject of random surveillance when you walk out in public. 

You might like the possibility that visual surveillance and biometric face recognition might be of assistance — sometimes — to law enforcement in doing their job, but there are times when you might wish to remain anonymous, even when you are engaged in activities that aren’t (yet) against the law.

If you are not a middle-aged white man, you should be more concerned with the possibility that one day you might be incorrectly identified as the perpetrator of a crime by faulty face recognition technology.

It’s an unfortunate fact that privacy in this country might as well take a bed in a hospice, as surveillance and identification technologies intrude on our daily lives, invade our public spaces, and impede our right to privacy.

George Orwell wrote 1984, a story of a world under constant surveillance, where every citizen was monitored, and couldn’t opt out from it. Forty years after that title, what Orwell wrote about has come to pass, and privacy is largely becoming a thing of the past.

There will no doubt be improvements to the hardware and software that run biometric facial recognition technologies. We should hope that mistakes in identity will happen less often.

But no matter how improved the tool gets, what will likely never be repaired is the past right to privacy we once enjoyed. Not to mention the relative freedom from law enforcement surveillance we once had.

[1] A BRIEF HISTORY OF FACIAL RECOGNITION, NEC  https://www.nec.co.nz/market-leadership/publications-media/a-brief-history-of-facial-recognition/

[2] FACIAL RECOGNITION TECHNOLOGY, GAO 3/27/24 https://www.gao.gov/products/gao-24-107372

[3] IBID.

[4] Public more likely to see facial recognition use by police as good, rather than bad for society 3/17/22, Raine, et. al. PEW https://www.pewresearch.org/internet/2022/03/17/public-more-likely-to-see-facial-recognition-use-by-police-as-good-rather-than-bad-for-society/

[5] IBID

[6] Op. Cit.

[7] BIASED TECHNOLOGY THE AUTOMATED DISCRIMINATION OF FACIAL RECOGNITION 2/29/24 Rachel Fergus https://www.aclu-mn.org/en/news/biased-technology-automated-discrimination-facial-recognition

[8] IBID.