Florida Man Wrongfully Arrested After Faulty Face-Recognition Match

Robert Dillon, a 52-year-old commercial crabber from Fort Myers, was arrested on suspicion of attempting to lure a child. Police relied on a match from the FACES facial-recognition system operated by the Pinellas County Sheriff’s Office. The system generated a 93 percent similarity score between Dillon’s face and a cell-phone photo of a suspect. Despite living more than 300 miles from the alleged crime scene and asserting he had never been to the city, Dillon was taken into custody at his home, held overnight in a cold cell, and transported in a caged, unlit van.

Wired Magazine

FACES maintains tens of millions of Florida mugshots and driver-license photos. It is among the longest-running police facial-recognition databases in the United States. The algorithm’s scores indicate how closely two images resemble each other, not the probability that they depict the same individual.

Wired Magazine

• Robert Dillon – the arrested individual.

Wired Magazine

• Pinellas County Sheriff’s Office – operator of the FACES system.

Wired Magazine

• American Civil Liberties Union (ACLU) – filed a lawsuit challenging the arrest and the system’s use.

Wired Magazine

The system returned a “93 percent match on facial features.” Scores reflect similarity, not identity certainty. The database contains tens of millions of images. Dillon’s residence is over 300 miles from the alleged incident. The arrest occurred during peak stone-crab season, causing Dillon to fall behind on rent and nearly lose his home.

Wired Magazine

Police investigative notes recorded the 93 percent match as the basis for the arrest. The ACLU’s complaint details that Dillon was detained overnight, required to pledge his truck’s title for bond, and that his mugshot remained online for nearly a year until a television reporter prompted its removal.

Wired Magazine

The ACLU argues that reliance on similarity scores misleads investigators and endangers civil liberties. Critics highlight that the algorithm’s output can be misinterpreted as definitive identification, leading to wrongful arrests and severe personal consequences, as illustrated by Dillon’s experience.

Wired Magazine

• “93 percent match on facial features,” — Police investigative notes

Wired Magazine

• “The scores it emits represent how much two images look alike to the algorithm.” — Police investigative notes

Wired Magazine

• “He pledged the title to his truck to make bond.” — ACLU complaint

Wired Magazine

• “The American Civil Liberties Union, which filed the suit, says Dillon was arrested at his home in front of his wife, held overnight in a cold cell, and transported in a caged, unlit van.” — ACLU complaint

Wired Magazine

Dillon’s arrest resulted in financial hardship, public scrutiny, and prolonged online exposure of his mugshot. The case underscores broader concerns about the accuracy of entrenched facial-recognition tools and their potential to infringe on individual rights.

Wired Magazine

The lawsuit remains pending, and its outcome may prompt review of the FACES system’s deployment, possible policy revisions, and increased oversight of facial-recognition technology in law-enforcement practices.

Wired Magazine