Florida Man Files Federal Lawsuit Following Wrongful Arrest Based on Flawed Facial Recognition Technology

Robert Dillon, a 52-year-old commercial fisherman from Fort Myers, Florida, has filed a federal lawsuit against multiple law enforcement agencies after a facial recognition error led to his wrongful arrest for a crime committed hundreds of miles from his home. The lawsuit, filed on Wednesday by the American Civil Liberties Union (ACLU), alleges that police investigators relied almost exclusively on an inaccurate algorithmic match while ignoring exculpatory evidence that would have cleared Dillon of any involvement in the attempted luring of a child.

The incident highlights growing concerns regarding the reliability of facial recognition technology in policing and the potential for digital "matches" to bypass traditional investigative rigor. Despite never having set foot in the city where the crime occurred, Dillon was treated as a primary suspect based on a 93 percent "similarity score" generated by a statewide database.

The Incident and the Algorithmic Match

The case originated from an incident shortly before midnight on November 2, 2023, at a McDonald’s in Jacksonville Beach, Florida. According to police reports, an unidentified male approached a girl under the age of 12 and repeatedly pressured her to leave the premises with him. The child refused and alerted her mother, prompting the suspect to flee before officers arrived.

During the initial investigation, a Jacksonville Beach police officer obtained photos of the suspect from the restaurant’s surveillance footage, captured via a cell phone camera pointed at a monitor. This low-resolution image was distributed in an "attempt-to-identify" bulletin. A sergeant with the Jacksonville Sheriff’s Office (JSO) processed the image through FACES (Facial Analysis, Comparison, and Evaluation System), a massive biometric database operated by the Pinellas County Sheriff’s Office.

The FACES system returned a result identifying Robert Dillon as a "93 percent match." However, as the lawsuit emphasizes, these percentages represent a similarity score—how much two images resemble one another according to an algorithm—rather than a statistical probability that the two images depict the same individual. Despite the inherent limitations of this technology, the match became the cornerstone of the state’s case against Dillon.

A Chronology of Investigative Failures

The legal complaint outlines a timeline of events that suggests law enforcement ignored several key facts that contradicted the facial recognition result.

In late November 2023, shortly after the crime, investigators spoke with a manager at the McDonald’s where the incident occurred. The manager explicitly stated that the suspect was a "regular customer" whom she had seen at the location multiple times. Dillon, a resident of Fort Myers, lived more than 300 miles away and maintained that he had no ties to Jacksonville Beach.

Furthermore, the investigating officer requested a search of automated license plate readers (LPRs) for two vehicles registered to Dillon. The search was intended to see if his truck or car had been in the Jacksonville area during the time of the crime. The results were negative; neither vehicle was detected anywhere in the county. According to the ACLU’s complaint, these findings—the manager’s statement and the negative LPR results—were omitted from the probable cause affidavit submitted to the judge who eventually signed the arrest warrant.

Following the initial flurry of activity in November 2023, the case reportedly went cold for six months. It was not until July 2024 that the investigating officer submitted the warrant application. A judge, unaware of the conflicting evidence, signed the warrant, leading to Dillon’s arrest in August 2024.

The Human and Financial Cost of Misidentification

For Robert Dillon, the consequences of the technical error were immediate and devastating. He was arrested at his home in Fort Myers in front of his wife, a traumatic event that he says has left lasting emotional scars. He was held overnight in a cold cell and transported across the state in a caged, unlit van.

To secure his release on bond, Dillon was forced to pledge the title to his commercial fishing truck. The timing of the arrest was particularly damaging, occurring during the peak of the stone crab season. As a commercial crabber, Dillon relies on this window for the bulk of his annual income. The lost time at sea and the legal expenses incurred led him to fall behind on his rent, nearly resulting in the loss of his family home.

Even after the legal system acknowledged the error, the stigma persisted. Dillon’s mugshot remained on the county’s public website for nearly a year, only being removed after a local television reporter intervened. The lawsuit notes that Dillon is frequently approached by strangers in public who recognize him from the arrest records, and he has expressed a profound loss of comfort in his daily life, particularly feeling uneasy when interacting with children in public settings due to the nature of the false charges.

The FACES Database: History and Oversight Concerns

The system at the heart of this controversy, FACES, has been in operation since 2001. Managed by the Pinellas County Sheriff’s Office, it is one of the oldest and most expansive facial recognition networks in the United States. It contains tens of millions of images, including Florida mugshots and Department of Highway Safety and Motor Vehicles (DHSMV) driver’s license photos.

At its peak, the system was accessible to more than 260 local, state, and federal agencies, including the FBI and Immigration and Customs Enforcement (ICE). Despite its widespread use, critics have long pointed to a lack of oversight. A 2016 study by Georgetown Law’s Center on Privacy and Technology, titled "The Perpetual Line-Up," found that the Pinellas County Sheriff’s Office conducted virtually no audits to monitor how the database was being searched and did not require officers to have "reasonable suspicion" before running a query.

When questioned about the lack of audits in the past, Pinellas County Sheriff Bob Gualtieri reportedly stated that such oversight was not a priority. This lack of transparency has raised alarms among civil liberties advocates, who argue that the technology is often used without the safeguards necessary to protect the constitutional rights of citizens.

A Systemic Pattern in Jacksonville

The ACLU asserts that Dillon’s case is not an isolated incident but part of a troubling trend. According to the organization, there have been at least 15 documented cases of wrongful arrests in the U.S. directly attributed to facial recognition errors.

The Jacksonville Sheriff’s Office, specifically, has faced similar allegations recently. Earlier in 2024, a North Carolina man was wrongfully arrested in connection with an auto-theft investigation after FACES returned an 85 percent match. That individual spent nearly three months in jail before charges were dropped. By the time of his release, he had lost his employment, his residence, and custody of his children.

These recurring errors have led to calls for a complete overhaul of how Florida agencies utilize biometric data. Jacksonville Sheriff T.K. Waters has publicly stated that a facial recognition hit should never serve as the sole basis for probable cause. "If you came to me with a facial-recognition hit and that was your probable cause, I would probably kick you out of my office," Waters told local media following the dismissal of Dillon’s case. However, the lawsuit alleges that, in practice, the JSO and surrounding agencies continue to allow these matches to drive the direction of criminal investigations with insufficient corroboration.

Legal and Policy Implications

The lawsuit filed on behalf of Robert Dillon names the investigating officer, the JSO sergeant, and the sheriffs of Pinellas and Jacksonville in their official capacities. The suit seeks compensatory and punitive damages for the violation of Dillon’s Fourth Amendment rights, which protect citizens against unreasonable searches and seizures.

Beyond financial restitution, the ACLU is asking the court to mandate a comprehensive overhaul of facial recognition policies within these departments. Proposed changes include:

• Mandatory Corroboration: Prohibiting arrests based solely or primarily on an algorithmic match.
• Transparency in Warrants: Requiring officers to disclose the use of facial recognition and any conflicting evidence (such as LPR data or witness statements) to judges when seeking warrants.
• Regular Auditing: Implementing strict logs and third-party audits to ensure the technology is not being misused or relied upon as a shortcut for traditional detective work.

Nate Wessler, deputy director of the ACLU’s Speech, Privacy, and Technology Project, emphasized the gravity of the situation. "No one should lose their freedom or be scared to leave their house because an algorithm got it wrong," Wessler said. "Police across the country are on notice: unreliable face-recognition technology is hurting people, and we will keep fighting to hold them accountable."

Analysis of Technology in Modern Policing

The case of Robert Dillon serves as a cautionary tale for the integration of Artificial Intelligence (AI) and biometric analysis in the justice system. While facial recognition can be a powerful tool for generating leads, its shift from a "lead-generation tool" to a "primary evidence tool" represents a significant risk to due process.

Algorithms are trained on datasets that may contain inherent biases, often performing less accurately on people of color or in varying lighting conditions. In Dillon’s case, the use of a cell phone photo of a surveillance monitor added a layer of digital noise that likely skewed the results. When law enforcement treats a "93 percent similarity" as a definitive identification, they ignore the probabilistic nature of the software.

As this litigation moves forward, it will likely prompt a broader legislative debate in Florida and across the nation regarding the "Black Box" nature of police technology. Without clear statutory requirements for how these systems are used and what must be disclosed to the defense during discovery, the risk of misidentification remains a persistent threat to the liberty of individuals who happen to share a passing resemblance to a digital ghost.

For Robert Dillon, the legal battle is about more than just a settlement; it is an effort to ensure that what happened to him—a life upended by a computer error and a lack of investigative follow-through—does not happen to another citizen. "I’m still picking up the pieces of my life," Dillon said, "all because the police relied on this dangerous technology instead of doing their jobs."