Anti Facial Recognition Case: 168 Flagged, 1 Jailed 11 Hours

Jason Killinger walked out of the Peppermill casino in Reno with nothing on his mind except getting home. He had his Nevada driver's license in his wallet, a player's card with his own name on it, a debit card, a UPS pay stub, his vehicle registration, and a union card. Nine days later, police showed up and put him in handcuffs anyway — because a facial recognition system, a computer program that compares your face to a database of photos, told them he was someone else entirely. He spent 11 hours in jail. None of the paperwork in his pocket mattered once the machine said "100% match." This is exactly the kind of case the anti facial recognition movement has been warning about for years, and now there's a number attached to it that's hard to look away from: 168.

TL;DR: A Reno man says he was wrongly arrested after facial recognition wrongly flagged him, and he's now demanding the names of 167 other people caught by the same system — a number that suggests this isn't a one-off glitch, it's a pattern nobody's been tracking.

Inside the Reno Casino Facial Recognition Case

Killinger, a truck driver, is now suing — and he's added the City of Reno itself as a defendant. According to Focus Gaming News, a federal judge ruled in April that his claims could move forward on the theory that Reno never trained its officers on what facial recognition can and can't actually prove. That's the part that should stop you mid-scroll. This wasn't a rogue cop making a bad call on his own. It was a system the department trusted, deployed without anyone apparently telling the officers, "hey, this thing gets it wrong sometimes — here's how often, and here's what to check before you cuff someone." This article is part of a series — start with How To Spot A Deepfake.

What happened during the Reno casino facial recognition arrest?

A casino security camera captured Killinger's face, and matching software linked it to a banned gambler. Despite Killinger showing valid ID and multiple documents proving his identity, police arrested him anyway, and he was held for 11 hours before the mistake came apart, according to reporting from Hoodline.

Here's where it gets interesting — and where Reno's lawyers are trying to draw a very careful line. According to ID Tech Wire, the city's attorney argues only 16 of the 168 flagged incidents happened before Killinger's arrest, so the rest can't prove the city already knew there was a problem. The casino, meanwhile, reportedly only handed over dates, times, and initials for the flagged people — not full names — citing privacy. Which is a strange thing to worry about privacy for, when you're the one who potentially got 167 other people mistaken for criminals.

Think about that math for a second. Even if you only count the 16 flags that happened before Killinger walked out in handcuffs, that's 16 separate moments where a computer told a security guard or a cop "this is your guy" — and it might have been wrong every single time. Killinger's lawsuit isn't really about one bad match anymore. It's about whether the city had every reason to know its system was throwing false positives (wrongly flagging innocent people as matches) long before it flagged him.

Rates of false positives are highest in East and West African and East Asian people, with a factor of 100 more false positives between different demographic groups. — Findings from a 2017 National Institute of Standards and Technology study of 140 facial recognition algorithms, cited via arXiv research analysis

The Anti Facial Recognition Movement Was Right About This

Let's be honest about what's actually being tested in this case. It's not whether facial recognition technology "works" in some general sense — cameras and software are genuinely good at narrowing down a list of possible matches. The real question is whether police, casinos, and cities are treating a computer's guess as if it were a fingerprint or a signed confession. It isn't. A face match is a lead, not proof. Killinger had six separate pieces of ID on him and it didn't matter, because the software said "100% match" and that number apparently outranked his entire wallet. Previously in this series: Facial Recognition Benefits.

Anti Facial Recognition Laws Are Spreading Fast

More than 20 cities — including Boston, San Francisco, and Pittsburgh — have already banned police departments from using facial recognition at all, according to the ACLU. And here's the detail that should end most arguments in this debate: no wrongful arrest tied to facial recognition has been reported in any city with an active ban. Not one. That's not a coincidence, and it's exactly why the anti facial recognition push keeps gaining ground in city councils that have actually looked at the data instead of the sales pitch.

The ACLU has documented at least 15 known wrongful arrests connected to facial recognition, and in nearly every single one, the person wrongly arrested was Black. That's not a rounding error. It lines up almost exactly with what the NIST research found — that these systems make more mistakes on Black faces, East Asian faces, and other groups that weren't well represented in the photos used to build the software in the first place. So when people say "the algorithm is biased," this is what they mean in plain terms: it was trained mostly on one kind of face, and it's worse at telling the rest of us apart.

What a Real Investigation Should Actually Look Like

If you've ever wondered whether a photo really shows what someone claims it does — a profile picture, a security still, a face in a crowd — that's the exact question this kind of technology exists to answer, and it's a fair thing to want checked. The problem isn't the tool. It's treating one number from one tool as the finish line instead of the starting point. A serious identity check compares a face across multiple photos, multiple angles, multiple contexts, and treats a strong match as a reason to look closer — not as a verdict. One useful thing you can actually do: if you're ever told a system "matched" you or someone you know with high confidence, ask specifically what else was checked besides that single photo. If the answer is "nothing," that's not an identification. That's a guess with good lighting. Up next: How To Spot A Deepfake 1 School Photo Is All It Takes.

Nobody's saying every facial recognition search ends in a wrongful arrest. Most probably don't. But that's precisely the problem with a number like 168 — nobody was counting until a truck driver's lawyer made them. State of Surveillance, an organization tracking these cases, has documented more than 13 dismissed wrongful arrest cases nationally, and in nearly every one, the person wrongly arrested was Black, according to State of Surveillance. Add that to the ACLU's count and you start to see the outline of something much bigger than one casino in Nevada.

Killinger's lawyers aren't just asking for names because they're curious. They want to know if 167 other people had their afternoon ruined the same way he did — pulled aside, questioned, maybe cuffed, all because a piece of software was confident and nobody was trained to treat that confidence as a reason to investigate rather than a reason to arrest.