Wrongful Arrest Exposes Failures in One of the Oldest Police Face-Recognition Tools in the US

Overview

A face-recognition program reported a 93 percent confidence match and sent Robert Dillon, a 52-year-old commercial crabber from Fort Myers, Florida, to jail for an attempted child abduction he had no link to. The ACLU and ACLU of Florida filed a federal lawsuit on June 10, 2026, against three police agencies, arguing officers treated a probabilistic software guess as a near-certain identification while ignoring evidence that cleared him. The case puts a spotlight on one of the oldest police face-recognition tools in the country, operated by the Pinellas County Sheriff's Office.

Key Takeaways

Robert Dillon was arrested in August 2024 at his Fort Myers home for a crime reported more than 300 miles away in Jacksonville Beach, a city the lawsuit says he had never visited.
A statewide face-recognition system run by the Pinellas County Sheriff's Office returned Dillon as a 93 percent confidence match from grainy surveillance images.
Police obtained an arrest warrant largely on the software result, the lawsuit says, while leaving out facts that pointed away from Dillon.
A license-plate-reader database showed none of Dillon's vehicles near Jacksonville Beach in the 48 hours around the incident, yet the case proceeded.
The ACLU counts Dillon among at least 14 people wrongly arrested in the US since 2019 after face-recognition misidentification, with one tally putting the figure at 15.
Even Jacksonville Sheriff T.K. Waters has said a face-recognition hit alone is not valid probable cause for an arrest.

Stats & Key Facts

#93 percent: the confidence level the face-recognition software reported for the false match to Dillon.
#300+ miles: the distance between Dillon's Fort Myers home and the Jacksonville Beach crime scene.
#At least 14: people the ACLU estimates have been wrongly arrested in the US since 2019 due to face-recognition errors.
#3: Florida police agencies named as defendants in the lawsuit.
#48 hours: the window the license-plate-reader search covered, finding no trace of Dillon's vehicles near the scene.
#52: Dillon's age at the time of the wrongful arrest.

Who Robert Dillon Is and How a Software Match Put Him in Jail

The man at the center of the case had no connection to the crime he was accused of.

Robert Dillon is a 52-year-old commercial crabber who lives in Fort Myers on Florida's southwest coast. In August 2024, deputies arrested him at his home and tied him to an attempted child abduction in Jacksonville Beach, a city on the opposite side of the state.

The link was not an eyewitness who knew him or physical evidence at the scene. It was a face-recognition algorithm that compared grainy surveillance photos to a database of images and reported Dillon as a 93 percent confidence match. The lawsuit says he had never set foot in Jacksonville Beach in his life.

The November 2023 McDonald's Incident That Started the Investigation

The case traces back to a report of a man trying to lure a young girl.

In November 2023, Jacksonville Beach police responded to a report that a man at a local McDonald's tried to get a girl under 12 to leave with him. Investigators pulled surveillance footage that showed the suspect, but the images were low quality.

Officers fed those images into a statewide face-recognition system operated by the Pinellas County Sheriff's Office. The lawsuit describes the suspect as a regular at the restaurant, a detail that pointed toward a local person rather than someone living five hours away.

How Police Treated a 93 Percent Guess as a Near-Certain ID

The core of the ACLU's complaint is misuse of a probabilistic tool.

›The system returned Dillon as a 93 percent match, a probability score, not a confirmed identification.
›Police obtained an arrest warrant on the strength of that software result, according to the lawsuit.
›Face-recognition vendors and many police policies say a match is only an investigative lead that needs independent corroboration.
›The ACLU argues officers across Florida are treating these scores as definitive proof rather than a starting point.

Evidence That Cleared Dillon Was Ignored or Left Out, Suit Says

The complaint accuses investigators of withholding facts that pointed away from Dillon.

Investigating officer Scott O'Connell queried a license-plate-reader database, which found no sign of Dillon's vehicles anywhere near Jacksonville Beach in the 48 hours around the incident. The lawsuit says this result, along with the 300-mile distance and Dillon's statement that he had never visited the area, was omitted from the warrant application.

Dillon's attorney later showed he was at work when the crime happened. Prosecutors then dropped the charges. Accounts of the timeline differ: one report says charges fell about two months after his arrest, while the ACLU describes more than a year of prosecution before the case ended.

The Three Agencies and Officers Named in the Lawsuit

The suit reaches beyond a single department.

›Jacksonville Beach Police Department, which handled the original abduction report.
›Jacksonville Sheriff's Office, whose employee ran the face-recognition search.
›Pinellas County Sheriff's Office, which operates the statewide face-recognition system.
›Individual officers involved in the investigation, including Scott O'Connell.
›The system is described as one of the oldest police face-recognition tools of its kind in the country.

A Growing Pattern of Wrongful Arrests From Face Recognition

Dillon's case is not isolated.

The ACLU estimates Dillon is one of at least 14 people wrongly arrested in the United States since 2019 after face-recognition technology misidentified them, with one tally placing the count at 15. Civil-rights groups say the cases share a pattern of departments leaning on automated matches as if they were certain.

Jacksonville Sheriff T.K. Waters has himself said that if an officer brought him a face-recognition hit as their probable cause, he would send them away, because that is not how the tool is meant to work. ACLU attorney Nate Freed Wessler said no one should lose their freedom because an algorithm got it wrong, and ACLU of Florida attorney Nicholas Warren said one wrongful arrest is one too many.

Frequently Asked Questions

What is Robert Dillon suing the police for?

He alleges that three Florida police agencies wrongfully arrested him based on a faulty face-recognition match and concealed evidence that showed he could not have committed the crime. The ACLU and ACLU of Florida filed the federal suit on his behalf on June 10, 2026.

What does a 93 percent face-recognition match actually mean?

It is a probability score the software assigns to how closely two images resemble each other, not a confirmed identity. Vendors and most police guidance treat such a match as an investigative lead that needs independent evidence before any arrest.

Why is the distance between the cities important?

Dillon lives in Fort Myers, more than 300 miles from Jacksonville Beach, and the lawsuit says he had never been there. A license-plate-reader search also found no trace of his vehicles near the scene, undercutting the claim that he was the suspect.

How many people have been wrongly arrested because of face recognition?

The ACLU estimates at least 14 people in the US since 2019 have been wrongly arrested after face-recognition misidentification, with one count putting the figure at 15. Dillon's case adds to that documented pattern.

Did the charges against Dillon get dropped?

Yes. Prosecutors dropped the charges after his attorney showed he was at work when the crime occurred. Reports differ on whether that took about two months or more than a year.

The lawsuit asks a court to examine how Florida police rely on automated face matching, a practice under rising scrutiny as documented wrongful arrests pile up. For Dillon, the case is about being jailed for a crime that happened hundreds of miles from a city he says he never visited.