CaraComp
CaraComp
Forensic-Grade AI Face Recognition for:
Get Started7-day refund guarantee**
facial-recognitionBy Cara Candelario

Facial Recognition False Arrest: Fargo Police Face $10M Suit

facial recognition false arrest, false facial recognition match, bold text graphic reading one false match one arrest
A bold graphic about a false facial recognition match leading to one arrest, showing why police need independent evidence. Illustration: CaraComp

Quick answer

What is a facial recognition false arrest and how does it happen?

A facial recognition false arrest happens when software wrongly ranks an innocent person's face as a likely suspect and police treat that guess as certain. Guidance says a result is only a lead. Officers should confirm it with independent evidence, such as an alibi check or a witness, before arresting anyone.

Picture a woman in Tennessee who gets arrested for a crime she says she had nothing to do with. The reason, according to her lawsuit, is that a computer said her face looked like someone else's. That is a facial recognition false arrest, and this one is now a $10 million lawsuit against Fargo police, who deny they broke any law.

A facial recognition false arrest shows why a computer's face match should be a lead to check, never proof that someone is the right person.

Let's be honest about what we know. This is one reported case. The coverage gives no error rate, so nobody should read it as "this happens X percent of the time." But one case is enough to ask a fair question: what should an officer have to do before a screen's guess turns into handcuffs?

Facial recognition false arrest: what happened and why it matters to ordinary people

Here is the short version, as reported by Biometric Update. A woman was arrested after a false facial recognition match. She filed a $10 million lawsuit. The police response was not "we're sorry." It was "we did nothing illegal."

That denial is the part worth sitting with. Police are not saying the match was right. They are saying that acting on it did not break the law. Those are two very different claims, and the gap between them is where ordinary people can fall.

What is a false facial recognition match, in plain English?

A false match (the software says two faces belong to the same person when they don't) is not a glitch in the movie sense. The system compares a photo of a suspect to a big pile of other photos. It then spits out the faces that look most alike. Somebody always ranks first, even if the real culprit isn't in the pile at all.

That is why the word "match" is a bit of a trap. It sounds like a verdict. It is really a ranked list of lookalikes.

How did one false match turn into an arrest?

According to the research behind this piece, the match came from another agency. A detective then used it to get a nationwide warrant. Think about what that does. Each step in the chain trusts the step before it, and nobody goes back to ask whether the first guess was any good. This article is part of a series, start with Fake Age Verification Steam Users Face Credit Card Scams.

It is a bit like a game of telephone, except the last person in line has a badge and a warrant.

8 of 8
known wrongful-arrest cases where police skipped at least one basic investigative step
Source: Washington Post analysis of eight wrongful arrest cases

That number comes from a Washington Post look at eight people wrongfully arrested after police used this technology. In all of them, police skipped something basic. In every known case, they arrested someone without independently tying that person to the crime. Not a fingerprint. Not a witness. Not even a check on where the person was that day.


Why a facial recognition false arrest sticks in the mind (and what the numbers really say)

Psychologists have a name for how our brains judge risk: the availability heuristic. Plainly, if a story is vivid and easy to picture, we assume it must be common. A woman pulled from her life over a computer's guess is exactly that kind of story.

So let's keep our heads. One case does not prove errors are everywhere. But the pile is growing. The ACLU has documented more than a dozen wrongful arrests tied to police reliance on this technology, and the research for this piece counts at least twelve false arrests in the United States. That is a real pattern, even if nobody can say how big it is.

If the software is accurate, why do false arrests still happen?

Fair question. The research cites top systems with false positive rates (how often the software wrongly says "same person") as low as 0.3 percent. Sounds tiny. But run a small error rate across millions of faces and real people get caught in it. And the same research notes that, in testing of leading systems, false positives were 23 times higher for West Africans than for Eastern Europeans, and 4.6 times higher for women than for men. The Bipartisan Policy Center explains how settings and demographics change those error rates.

Here's the real kicker, though. The error rate may not be the main problem. The main problem is what people do with the answer.

Facial recognition results are never "matches." Independent investigation is required.

Guidance from leading law enforcement agencies, as summarized in the research behind this article

Read that again. The law enforcement agencies that use this tool say, in their own guidance, that a result is not a match. Yet in the known wrongful arrest cases, somebody treated it like one anyway. Previously in this series: Id Scan How A 93 Face Match Put The Wrong Man In Cuffs.


Trusted by Investigators Worldwide
Run Forensic-Grade Comparisons in Seconds
Detailed facial comparison reports. Results in seconds.
Get Started
7-day refund guarantee**

Police facial recognition: a lead is not proof

Good detectives have always worked with leads. A tip, a blurry photo, a neighbor's hunch. None of those gets you an arrest by itself. A false match from a computer should be treated the same way, with one extra twist: it looks scientific, so it feels more trustworthy than a hunch. It isn't.

Here is the contrast that matters most.

StepTreating a match as a leadTreating a match as proof
First movePull the lookalike list and ask who else could fitTake the top name as the suspect
Independent evidenceRequired before any arrest (an alibi check, a witness, a record)Skipped
A match from another agencyRe-checked by the officer who acts on itPassed along and used for a warrant
If the face is wrongCaught early, nobody is arrestedA facial recognition false arrest

The left column costs a few hours of police work. The right column can cost a person their freedom, their job and their name. It also costs a city a lawsuit.

Can police be held responsible when they act on someone else's match?

That is the open question in this case. The research says it will help decide how legal accountability works when police act on leads from other agencies, especially when their own department has no formal policy or training on the tool. Most departments do have written rules telling officers to corroborate a match. But rules are guidance, not law. The police denial suggests they believe following their own process shields them, even if the process was thin.

My opinion? If a written policy says "never treat a result as a match," then ignoring it should carry weight in court. Otherwise the policy is a poster on a wall.

Why a facial recognition false arrest matters to you

  • You don't have to do anything wrong - a lookalike is enough to put a person under suspicion.
  • The chain hides the weak link - once a match is passed between agencies, nobody re-checks the first guess.
  • Policy without law is optional - rules that say "verify first" mean little if breaking them has no consequences.

What should be checked before a face match leads to an arrest?

Before an officer acts on a face match, there should be at least one piece of evidence that has nothing to do with the face. Where was the person that day? Does a witness pick them out? Does anything physical connect them to the scene? If the only thing pointing at them is a computer's ranking, that is a reason to keep investigating. It is not a reason to arrest.

If you've ever wondered whether a photo or profile is really who it claims to be, that is the exact question this kind of technology exists to answer. Comparing two faces side by side can be genuinely useful. The one thing worth carrying with you is this: a good result is the start of the check, not the end of it. Ask what second, independent clue backs it up. If the answer is "nothing," treat the match as a maybe.

Key Takeaway

A facial recognition false arrest happens when a lead gets treated as proof, so a face match should always be backed by independent evidence before anyone is arrested. Up next: Facial Recognition False Arrest Fargo Cops Face 10m Suit.

Recap: A computer's face match is only a lead for police to check. Without other evidence behind it, acting on it can wrongly put an innocent person under arrest.

So here is the question I can't shake. If the rule everyone agrees on is "a result is never a match," and a woman was arrested anyway, what exactly is the rule for? The next facial recognition false arrest will not announce itself. It will look like a normal Tuesday, a name at the top of a list, and nobody asking "who else could this be?"

Facial recognition false arrest: Frequently Asked Questions

What is a facial recognition false arrest?

It is an arrest that happens because facial recognition software wrongly says a person's face matches a suspect's. The person is innocent, but the computer's suggestion is treated as if it were proof. The reported case here involves a woman, a false match, a $10 million lawsuit and police who deny legal wrongdoing. Experts say the arrest usually follows when officers skip basic checks, such as confirming where the person was or finding other evidence that links them to the crime.

How often does police facial recognition get it wrong?

This story does not give an error rate, so no one should claim one from it. It is one reported case. The wider research cites at least twelve false arrests in the United States and one figure of a 0.3 percent false positive rate for top systems. Even a small rate can touch real people when many faces are searched. Error rates also differ by group, according to testing the research describes, which is why independent checking matters so much.

Is a face match enough evidence to arrest someone?

Law enforcement guidance quoted in the research says no. It says results are never matches and that independent investigation is required. A face match is best treated as a lead, like a tip. Before an arrest, officers should find evidence that does not rely on the software, such as an alibi check, a witness or a physical link to the crime. In the known wrongful arrest cases, that step was missing.

Can police be sued over a false facial recognition match?

The woman in this case filed a $10 million lawsuit, and police say they did nothing illegal. How the courts rule could shape how accountability works when officers act on a match from another agency. Many departments have written policies telling officers to verify a match first, but policies are guidance and not law. Whether ignoring one counts as a legal failure is the open question here. Anyone affected can ask an attorney about their options; this is general information, not legal advice.

What can an ordinary person do about police facial recognition?

You can't control a police database, but you can ask smart questions. Does your local department have a written policy on facial recognition technology? Does it require evidence beyond the match before an arrest? Is that rule backed by training? Asking at a city council meeting or in a records request puts the issue on the table. And if you ever compare faces yourself, treat the result as a clue to confirm with a second, independent source.

Ready for forensic-grade facial comparison?

Full forensic reports with detailed similarity scoring. Results in seconds.

Run My First Search