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

A truck driver walked into a Reno casino with a wallet full of I.D., his Nevada license, his player card, his union card, even a pay stub. None of it mattered. A facial recognition system said he was a banned patron with one hundred percent confidence, and he spent eleven hours in a cell.

His name is Jason Killinger

His name is Jason Killinger. Police say a camera matched his face to someone the casino had banned. He handed over every piece of identification he was carrying. It didn't stop the arrest. And now he's asking a question that should stop all of us cold, how many other people did that same system flag? The answer, according to his lawsuit, is a hundred and sixty-eight. If you've ever had your face caught on a camera in a store, an airport, or a casino, this story is about you. So here's what we're chasing today, should a computer's guess be enough to put a person in handcuffs?

Start with what was in Killinger's pockets. A Nevada I.D. A Peppermill player card. A debit card. A U.P.S. pay stub. Vehicle registration. A union card. That's a stack of proof that he was who he said he was. The system said otherwise, and the humans trusted the system over the paper in his hand. That's the part that should sit with you. A machine's opinion outweighed six forms of I.D.

Now the number that started the fight. His lawyers want the names of the other hundred and sixty-eight people the same system flagged. Why? Because if that many people got caught in the net, one of two things is true. Either the system is spitting out a flood of false matches, or the casino and the city knew about a problem and did nothing. For a lawyer, that's about proving a pattern. For everyone else, it means you could be number a hundred and sixty-nine and never know it.

The casino says it only handed over dates, times,

The casino says it only handed over dates, times, and initials, not full names, citing privacy. The city argues most of those flagged cases came after Killinger's arrest, so they couldn't have known. But that argument steps on a landmine. If a hundred and sixty-eight people tripped the same alarm, the volume alone tells you something's broken.

And this isn't one glitch in one town. The American Civil Liberties Union has documented at least fifteen wrongful arrests tied to facial recognition. In nearly every single one, the person arrested was Black. That's not bad luck. Back in 01/01/2017, the National Institute of Standards and Technology tested a hundred and forty of these algorithms. They found the machines were far more likely to falsely match people with East Asian, West African, and East African faces, in some cases a hundred times more often. So the odds of being wrongly flagged aren't the same for everyone.

Here's the piece that changes how you see all of it. More than twenty cities, Boston, San Francisco, Pittsburgh, have banned police from using this technology. And in every city with an active ban, not one wrongful arrest has been reported.