Florida man sues police after facial recognition error led to wrongful arrest

When artificial intelligence systems are entrusted with critical decisions, the consequences can be severe—and in some cases, catastrophic. A recent lawsuit filed in Florida underscores this danger, alleging that police officers arrested an innocent man after a facial recognition algorithm produced a questionable match and then failed to pursue evidence that could have cleared him.

Robert Dillon, a 52-year-old resident of Fort Myers, was arrested in August 2024 on charges of attempting to lure a child near a Jacksonville Beach McDonald’s. The accusation stemmed from a surveillance video that prosecutors claimed showed Dillon approaching a child. However, the lawsuit argues that the entire case was built on a flawed AI analysis that police treated as definitive, despite glaring inconsistencies.

A 93% match with no corroborating proof

The lawsuit states that Jacksonville Beach police used a facial recognition system to scan the McDonald’s surveillance footage and identified Dillon as a 93% match to the suspect. The image used for the match was of particularly poor quality—essentially a photo of a computer screen displaying the grainy surveillance footage. This low-resolution source raised immediate red flags about the reliability of the identification, yet officers proceeded with the arrest rather than conducting further verification.

Dillon’s legal team argues that the arrest was the result of confirmation bias, where law enforcement prioritized the AI’s output over evidence that contradicted it. "This case is about what happens when police let an error-prone artificial intelligence system stand in for an investigation," the lawsuit states. "A facial recognition algorithm flagged Robert Dillon as the man who tried to lure or entice a child under twelve years old at a Jacksonville Beach McDonald’s. It was wrong."

Geographic and digital alibis ignored

The lawsuit outlines multiple ways Dillon could have been ruled out as a suspect, yet these were dismissed or overlooked. Dillon lived more than 300 miles from Jacksonville Beach, a fact that should have raised immediate doubts about his involvement. Additionally, a search of license plate reader databases in the area found no records of Dillon’s vehicle being present during the alleged incident. These digital breadcrumbs—readily available to investigators—were never pursued.

Instead of questioning the AI’s output, the lawsuit claims officers constructed a case designed to support the flawed identification. Dillon was arrested, prosecuted, and publicly stigmatized for one of the most damaging accusations an individual can face. The emotional and reputational toll of such an ordeal is incalculable, particularly when the initial evidence was so flimsy.

The broader implications of AI in policing

This case is not an isolated incident but a symptom of a growing trend: the uncritical adoption of AI tools in law enforcement without adequate safeguards. Facial recognition technology, while marketed as a precise investigative tool, has repeatedly demonstrated vulnerabilities to false positives, especially when used with low-quality or incomplete data. Critics argue that over-reliance on such systems can lead to wrongful convictions, racial biases, and a erosion of public trust in policing.

Legal experts warn that cases like Dillon’s could become more common as AI tools are integrated into investigative processes without proper oversight. "When algorithms are treated as infallible, the human element of justice is compromised," said one civil rights attorney familiar with the case. "Evidence must be scrutinized, not rubber-stamped by a machine."

The lawsuit seeks damages for wrongful arrest, emotional distress, and reputational harm, while also demanding reforms to prevent similar miscarriages of justice in the future. As AI continues to permeate law enforcement, the Dillon case serves as a stark reminder: technology should assist investigations, not dictate their outcomes.