Reno Cop Admits AI Arrest "Never Should Have Happened"

TL;DR: Jason Killinger spent 11 hours in a Reno jail (four of them handcuffed) after the Peppermill Casino’s facial recognition flagged him as a trespasser named Michael Ellis. Problem: Killinger had a valid Nevada driver’s license, UPS pay stub, and vehicle registration all proving exactly who he was. On January 22, 2026, Officer Richard Jager admitted under oath that the arrest “never should have happened.” But the lawsuit goes further: it alleges Jager “knowingly inserted false statements” into police reports claiming Killinger’s IDs were fraudulent. Trial is expected in 2026.

September 17, 2023: The AI Said 100%

Jason Killinger pulled into the Peppermill Casino in Reno. He was a long-haul UPS driver, a regular gambler, and he had no idea his face was about to betray him.

The casino’s facial recognition system scanned his face and returned a “100% match” to Michael Ellis, a man who’d been banned months earlier for sleeping on the premises.^[1]

Security detained Killinger. They called the cops. Officer Richard Jager arrived.

Here’s what Killinger handed over:

A valid Nevada driver’s license with his name and photo
A UPS pay stub confirming his employment
Vehicle registration matching his identification

All of it matched. Name. Photo. Physical description. The man standing in front of Jager was Jason Killinger.

But the AI said 100%. And that was enough.

The Bodycam Tells the Story

Released bodycam footage captures Officer Jager comparing Killinger’s ID to the casino’s records for the banned patron “Michael.”^[2]

The physical differences were obvious:

Height: Four-inch difference between the two men
Eye color: Killinger has blue eyes. Ellis has hazel.
Facial features: Beyond superficial resemblance, distinct differences

None of that mattered. On the bodycam, Jager is heard saying: “So the software’s saying it, it’s legit,” followed by “Alrighty. Well, that solves that.”^[2]

He also told a casino manager that Killinger “probably has a DMV hook-up”: suggesting fraudulent ID access, with zero evidence.^[2]

Killinger was arrested. Cuffed. Booked.

11 Hours in Custody

Killinger sat in the Washoe County Jail for 11 hours. Four of those hours he spent in handcuffs, sustaining bruising and shoulder pain.^[3]

Finally, someone ran his fingerprints.

The fingerprints confirmed what his driver’s license, pay stub, and vehicle registration had shown all along: he was Jason Killinger, not Michael Ellis.

He was released. The charges were dropped. The AI had been wrong.

But according to the lawsuit filed in July 2025, the story doesn’t end there.

The Lawsuit: Fabricated Evidence

Killinger filed a federal civil rights lawsuit in U.S. District Court for the District of Nevada (Case No. 3:25-cv-00388) on July 30, 2025.^[4]

The allegations against Officer Jager go beyond a simple mistake:

Fabrication of Evidence

Jager “knowingly inserted false statements” into both the incident report and arrest declaration, claiming Killinger’s IDs were “conflicting” and “appeared to be fraudulent.”^[4]

Omitting Exculpatory Evidence

The fingerprint results proving Killinger’s identity were allegedly excluded from police reports.^[4]

Malicious Prosecution

Wrongful initiation of trespassing charges against someone with clear proof of identity.^[4]

Constitutional Violations

Fourth and Fourteenth Amendment claims for fabricated evidence and due process deprivation.^[4]

Killinger’s attorney, Terri Keyser-Cooper, argues the city of Reno failed to properly train officers on facial recognition limitations.^[3]

January 2026: The Officer Admits It

On January 22, 2026, Jager sat for a deposition. Under oath, he made a remarkable admission.^[3]

“[The arrest] never should have happened.”

Jager said he now understood facial recognition should be treated as “an investigative lead only,” requiring corroboration before arrest. He acknowledged that a wrongful arrest is “a terrible thing.”^[3]

The kicker: Jager said this training came after the lawsuit was filed. The Reno Police Department apparently didn’t tell officers that AI matches aren’t probable cause until someone sued them over it.

Peppermill Settled. The City Fights On.

The Peppermill Casino quietly settled with Killinger for an undisclosed amount.^[3] The casino deployed the facial recognition system. The casino’s security detained him. The casino paid.

The city of Reno is still fighting. In March 2026, the city attorney’s office filed court documents alleging that Killinger and local news outlet This Is Reno “conspired” to influence the case outcome.^[5]

The trial is expected later in 2026.

This Keeps Happening

Killinger joins a growing list of people wrongfully arrested after facial recognition misidentified them:

Robert Williams (Detroit, 2020): Arrested at home in front of his daughters. AI matched him to a shoplifting suspect. Wrong guy.
Porcha Woodruff (Detroit, 2023): Seven months pregnant. Arrested for carjacking. AI match. Wrong person.
Trevis Williams (NYPD, 2025): Jailed for two days. He was 8 inches shorter and 70 pounds lighter than the actual suspect. Only similarity: both Black men with locks.

A Washington Post investigation found at least 15 police departments in 12 states arrested suspects based solely on AI matches, with no corroborating evidence.^[6] Of the eight confirmed wrongful arrests documented, seven victims were Black.

Killinger is white. The algorithm doesn’t just fail Black people. It fails everyone. The difference is how often.

What “100% Match” Actually Means

The Peppermill’s system returned a “100% match.” That sounds definitive. It isn’t.

Facial recognition systems generate probability scores, not fingerprints. A high confidence score means the algorithm thinks the faces are similar. It doesn’t mean they’re the same person. Different lighting. Different angles. Different ages. Any of these can throw off the match.

The FBI’s own facial recognition system explicitly warns that results should only be treated as investigative leads. Yet Officer Jager heard “100%” and said, “Well, that solves that.”

Every facial recognition vendor includes disclaimers in their terms of service: don’t use this as sole basis for identification. Police ignore them anyway.

The Training Gap

According to Jager’s deposition, the Reno Police Department only clarified that facial recognition isn’t sufficient for probable cause after Killinger sued.^[3]

That means every arrest based on AI matching before July 2025 happened without officers understanding the technology’s limitations. How many other people did Reno PD arrest on facial recognition alone? How many cases were never challenged?

The lawsuit argues this is a systemic failure, one that makes the city of Reno liable, not just the individual officer.

If This Happens to You

Document Everything

Keep receipts, photos with timestamps, location history. If you can prove you were elsewhere, the AI match falls apart.

Demand to Know

Ask if facial recognition was used in your case. Request all documents related to the match, including confidence scores and the comparison images.

Call a Lawyer

Civil rights attorneys and the ACLU are tracking these cases. Wrongful arrests based on AI matches may support Section 1983 civil rights claims.

Request Fingerprints

Fingerprints are what cleared Killinger. If you’re detained based on facial recognition, demand fingerprint verification before booking.

The Bottom Line

Jason Killinger had three pieces of government-issued identification proving who he was. The AI said 100%. The cop said “that solves that.”

Killinger spent 11 hours in jail. The fingerprints cleared him. The officer now admits under oath it “never should have happened.”

But the lawsuit says more happened. It alleges the officer fabricated evidence, called valid IDs fraudulent, and omitted exculpatory fingerprint results from reports. The casino paid. The city is still fighting.

Trial is expected in 2026. One question hangs over the case: if the AI said 100% and the cop said “that solves that,” what’s your ID actually worth?