ACLU Sues Florida Police Over Wrongful Arrest Tied To AI Match
The ACLU filed a lawsuit on Wednesday, June 10, 2026, on behalf of Richard Dillon against the Jacksonville Beach Police Department and others over an alleged wrongful arrest tied to an AI facial-recognition match.[1]
Dillon says he was wrongly accused of trying to lure a 12-year-old from a Jacksonville Beach McDonald's in November 2023 and that he was more than 300 miles away at the time.[1] Police submitted cellphone photos of surveillance stills to the FACESNXT system, which produced a reported 93 percent match and led to Dillon's arrest in August 2024 and an overnight jail stay.[1] Charges were later dropped after the state attorney's office reviewed the case, the lawsuit says.[1]
In November 2023, a man attempted to lure a child from a McDonald's in Jacksonville Beach.[1] Police months later sent still images to the Pinellas County Sheriff's Office's FACESNXT system, which Jacksonville officers accessed to run a match.[1] The suit says officers obtained an arrest warrant without confirming the AI result despite alibi evidence, including the absence of Dillon's vehicle in automatic license plate reader data.[1] The National Institute of Standards and Technology's Face Recognition Vendor Testing from 2018-2019 found that even the most accurate algorithms had identification error rates above 20 percent for low-quality images like surveillance photos.
At least 14 people in the United States are publicly known to have been wrongfully arrested after erroneous facial-recognition results, a tally current as of April 2026. Dillon said he is "still picking up the pieces" after the arrest, his lawyers told reporters as they filed the suit.[1]
The mainstream summary does not address the broader implications of Dillon's case in the context of systemic issues surrounding facial recognition technology. While it mentions that at least 14 people have been wrongfully arrested due to erroneous facial recognition results, it fails to highlight the significant error rates associated with these technologies, particularly in low-quality images like those from surveillance footage. According to the National Institute of Standards and Technology, even the most accurate algorithms had identification error rates exceeding 20% for such images, raising serious questions about their reliability in law enforcement contexts NIST.
Moreover, social media discussions reveal a growing concern about the implications of relying on flawed AI outputs in policing. Commentators have pointed out that the reliance on such technology can lead to a 'guilty until proven innocent' mentality and emphasize the need for traditional verification methods to safeguard civil liberties. This perspective suggests that the mainstream account may downplay the potential dangers of unregulated AI use in law enforcement, which could exacerbate existing biases and injustices in the criminal justice system AiquestAcademy.
Show source details & analysis (1 source)
📊 Relevant Data
NIST Face Recognition Vendor Testing found that even the most accurate algorithms had identification error rates above 20% for lower-quality images such as surveillance footage, webcam photos, or images captured from screens, compared to false negative rates as low as 0.25% for high-quality mugshot images.
Facial Recognition Technology (Part III): Ensuring Commercial Transparency and Accuracy — NIST
At least 14 people in the United States are publicly known to have been wrongfully arrested due to erroneous facial recognition results in police investigations.
More than a Dozen Wrongful Arrests Due to Police Reliance on Facial Recognition Technology — ACLU
📌 Key Facts
- On Wednesday, June 10, 2026, the ACLU filed a lawsuit for Richard Dillon against the Jacksonville Beach Police Department and others over an alleged wrongful AI-based arrest.
- Dillon was accused of trying to lure a 12-year-old from a Jacksonville Beach McDonald's in November 2023 but says he was more than 300 miles away; charges were later dropped.
- Police used the FACESNXT facial recognition system, which produced a reported 93% match from cellphone photos of surveillance footage, leading to Dillon's August 2024 arrest and overnight jailing.
📰 Source Timeline (1)
Follow how coverage of this story developed over time