The phone call on December 15, 2023, seemed so improbable that Robert Dillon thought it might be a scam.
According to a federal lawsuit filed on June 10, 2026, a Jacksonville Beach police investigator told the 52-year-old Fort Myers commercial crabber that he was being investigated in connection with an attempted child-luring case more than 300 miles away. Dillon, who said he had never been to Jacksonville Beach, immediately denied any involvement.
Eight months later, deputies arrived at his home and arrested him in front of his wife.
Dillon’s lawsuit alleges that investigators relied heavily on an AI-powered facial recognition search that incorrectly identified him as a suspect and failed to adequately pursue evidence that could have excluded him before his arrest. Prosecutors later dropped the charge.
His case is part of a growing number of wrongful-arrest claims across the country involving facial recognition technology. Similar lawsuits have surfaced in Michigan, New Jersey, Maryland, Georgia, Louisiana, and elsewhere, fueling debate over the role AI should play in criminal investigations and whether investigators sometimes place too much confidence in algorithmic results.
“Robert’s case illustrates the stakes when police deploy AI-assisted identification tools without adequate safeguards. Digital information can be a powerful tool for law enforcement, but its proliferation, supercharged by the AI boom, carries profound Fourth Amendment implications,” said Steve Silverberg, counsel at Hoguet Newman Regal & Kenney, LLP. “We are proud to stand with Robert and the ACLU in holding these agencies accountable, and in making clear that technological advancements, however enticing, do not suspend constitutional obligations.”
Mr. Dillon is represented by the American Civil Liberties Union, the ACLU of Florida, and Hoguet Newman Regal & Kenney, LLP.
Facial recognition systems are increasingly used by law enforcement agencies to compare surveillance images against databases containing millions of photographs. Supporters say the technology can generate leads quickly and help solve cases that might otherwise remain unsolved. Critics argue that the technology can create a false sense of certainty, particularly when poor-quality images are involved or when investigators treat a software-generated match as more than an investigative lead.
According to Dillon’s complaint, Jacksonville Beach investigators were attempting to identify a man captured on surveillance footage at a McDonald’s restaurant in November 2023. The lawsuit alleges that investigators relied on cellphone photographs taken of surveillance footage displayed on a computer screen rather than a high-quality digital image. Those photographs were later submitted to Florida’s Face Analysis Comparison and Examination System, or FACES, a facial recognition platform operated by the Pinellas County Sheriff’s Office.
The system returned Dillon as a “93% match,” according to the lawsuit.
A figure like 93% sounds compelling, but facial recognition experts have long cautioned that such scores do not represent the probability that a match is correct. Instead, they reflect the degree of similarity between mathematical templates generated by the software.
That distinction lies at the heart of a growing debate over facial recognition technology.
Most law enforcement agencies describe facial recognition as an investigative tool rather than evidence capable of establishing guilt. Jacksonville Beach Police Department policy cited in the lawsuit states that facial recognition results are intended to serve as investigative leads and not as positive identifications.
Yet critics argue that once a potential suspect is identified by software, investigators can become overly focused on that individual. Researchers refer to the phenomenon as “automation bias” — the tendency to place excessive trust in computer-generated outputs even when contradictory information exists.
The growing number of such cases has prompted some jurisdictions to ban law enforcement use of facial recognition technology, while others have imposed restrictions requiring additional oversight and corroborating evidence.
As AI tools become more sophisticated and more widely used, the debate is likely to intensify. Facial recognition technology can help investigators process vast amounts of information in a fraction of the time required by traditional methods. But critics argue that its value depends on understanding its limitations and treating its results as one piece of evidence rather than a conclusion.
