Exemplary damages, also called punitive damages, are money awarded to punish serious misconduct, not just to repay a loss. When a predictive-policing model points officers at the wrong person, the facts that can support that award show an officer knew the tool was unreliable and acted anyway, with "reckless or callous indifference" to the person's constitutional rights.
The software's mistake alone is rarely enough. What matters is what the officers knew, what rules they had to follow, and how they treated the person after the lead appeared. A damages claim built on those facts is also limited by who can be made to pay.
Table of Contents
- What Legal Standard Applies to Punitive Damages Against Police?
- Why the Algorithm's Error Is Not the Key Fact
- Which Real Cases Show What These Facts Look Like?
- Who Can Be Made to Pay Punitive Damages?
- What Records Make the Strongest Case?
- Frequently Asked Questions
What Legal Standard Applies to Punitive Damages Against Police?
Most claims over a false police lead are federal civil-rights suits under 42 U.S.C. § 1983. That law lets people sue state and local officials who violate their constitutional rights. In the Supreme Court's decision in *Smith v.
Wade* (1983), the Court held that a jury may award punitive damages when an official acted with evil motive or with "reckless or callous indifference" to federally protected rights. The plaintiff does not have to prove actual malice. That lower bar matters here. An officer who never meant to harm anyone can still face punitive damages. The question is whether the officer ignored a known risk to someone's rights.
Why the Algorithm's Error Is Not the Key Fact
The Court in *Smith v. Wade* noted that the common law, in 1871 and today, allows punitive damages for reckless indifference to other people's rights. Applied to a false lead, that puts the focus on human choices, not on the software.
A model that produces a bad lead is a faulty tool. An officer who knows the tool is unreliable and still arrests, searches, or targets someone on its word alone may have been reckless. The stronger claim rests on facts like these:.
- The officer knew, or was warned, that the tool often got things wrong.
- Department rules required backup evidence, and the officer skipped it.
- The officer acted on the lead alone, without independent checking.
- The officer ignored evidence pointing away from the person.
- The officer kept targeting the person after the lead proved weak.
Which Real Cases Show What These Facts Look Like?
Detroit police wrongfully arrested Robert Williams in 2020 at his Farmington Hills home, in front of his wife and two children, over a watch theft. According to the ACLU's case page on *Williams v. City of Detroit*, officers relied on a wrong facial-recognition match. It was the first publicly reported false-match arrest. Detroit settled on June 28, 2024. The city paid damages and attorneys' fees and barred arrests and photo lineups based only on a facial-recognition lead without independent, reliable evidence.
Rules like these define careful use, so a later officer who breaks similar rules has a harder time claiming an honest mistake. In Florida, the Pasco County Sheriff's Office ran an "intelligence-led policing" program that listed people it predicted would commit crimes and sent deputies to their homes. The Institute for Justice reports that the office admitted in a settlement that the program violated four residents' First, Fourth and Fourteenth Amendment rights. It agreed to pay $105,000 and never to run a similar program. One Pasco plaintiff, Darlene Deegan, faced three years of repeated home visits after the program flagged her son, as reported by the Florida Phoenix. She also ran up about $3,000 in minor code-violation fines, allegedly as retaliation for not cooperating. Repeated targeting and retaliation are the kind of conduct that can support a finding of callous indifference.
Who Can Be Made to Pay Punitive Damages?
This is the biggest limit on these claims. In *City of Newport v. Fact Concerts* (1981), the Supreme Court held that cities are immune from punitive damages under § 1983. It threw out a $200,000 punitive award against Newport, Rhode Island.
In practice, a plaintiff can seek ordinary compensation from a city or department, but exemplary damages must come from individual officers. That shapes the whole case. The evidence has to tie specific people to specific reckless decisions, not just describe a bad department-wide system. A claim that only blames "the program" or "the software" may recover compensation and still leave no one who can be ordered to pay punitive damages.
What Records Make the Strongest Case?
The ACLU says people of color and women are much more likely to be misidentified by facial recognition. Records showing that officers knew about those error rates, or broke department rules requiring corroboration, are the strongest facts against individual officers.
Useful records to request in discovery or through public-records laws include: Know the limits of the case record. The Pasco case was filed in federal court in 2021 and settled just as trial was about to start. No jury ruled on punitive damages in Pasco or in Williams, so neither case is a court precedent for exemplary damages over an algorithm's lead.
- Training materials and vendor warnings about the tool's accuracy
- Department policies on using algorithmic leads
- The officer's reports, notes, and body-camera footage
- Emails or messages about the person or the lead
- A log of visits, stops, citations, or fines tied to the person
Frequently Asked Questions
Can I get punitive damages from the police department itself?
Not under § 1983. *City of Newport v. Fact Concerts* holds that cities are immune from punitive damages, so those damages must be sought from individual officers.
Do I have to prove the officer wanted to hurt me?
No. *Smith v. Wade* allows punitive damages for reckless or callous indifference to your rights, without proof of actual malice.
Did Robert Williams or the Pasco plaintiffs win punitive damages at trial?
No. Both cases settled, so no jury decided punitive damages in either one.
You Might Also Like
- A Hiring Algorithm Rejecting Applicants Discriminatorily: What Facts Could Support an Exemplary-Damages Claim?
- A Deepfake Voice Used to Authorize a Bank Transfer: What Facts Could Support an Exemplary-Damages Claim?
- Severe Turbulence Causing Unrestrained Passenger Injuries: Which Economic and Non-Economic Damages Could Be Available?