deniz.in

Markets

Weather

Loading weather

· via Hacker News – Front Page (native)

Black students sue Omnilert and Baltimore County after AI flags Doritos bag as gun

Four Black students are suing the maker of an AI gun-detection system, their Maryland school district and county police after the software flagged a bag of chips as a firearm and armed officers detained them.

Black students sue Omnilert and Baltimore County after AI flags Doritos bag as gun

Four students sue over an armed response to a bag of chips

The Lawyers' Committee for Civil Rights Under Law has filed a civil rights lawsuit on behalf of four Black current and former high school students, arguing that an AI weapon-detection vendor, a Maryland school district and county police unlawfully detained and searched the teenagers after the software mistook a bag of chips for a firearm.

The defendants include Omnilert LLC, maker of Omnilert Gun Detect; Baltimore County Public Schools and its personnel; the Baltimore County Police Department; and individual officers. The case, Allen, et al. v. Omnilert, et al., was filed in the Circuit Court for Baltimore County, Maryland. The committee is joined by Willkie Farr & Gallagher LLP as pro bono counsel and by Engel Law Group.

What happened at Kenwood High School

According to the Lawyers' Committee, the incident took place in October 2025 at Kenwood High School in Baltimore County. Omnilert Gun Detect, which the district uses across its schools, generated an alert stating a 90 percent probability that a gun was on campus. The object in question was a teenager's bag of chips, identified in the Hacker News headline on the announcement as a bag of Doritos.

The committee describes the false alert as the start of a chain of human failures. School staff and local police reviewed the image, determined it was a mistake and cancelled the alert. Armed officers were dispatched anyway, because school and police officials lacked protocols to keep a cancelled alert from triggering an armed response. Officers then confronted the students just steps from the school entrance, ordered them to their knees, handcuffed and searched them, although they had done nothing wrong. One student's mother, Freida Winfield, was arriving to pick up her son when it happened; she later said no child in the county should face that kind of danger while waiting outside school for a ride home.

The committee also states that school officials knew the system regularly produced false alerts, and that the alert kept circulating even after staff and police had reviewed the image and recognized the error.

The claims

The complaint alleges violations of the teenagers' federal and state constitutional rights, plus negligence and gross negligence by the defendants. "Overreliance on artificial intelligence technology mixed with inept communication and coordination among key officials is a toxic mix," said Damon Hewitt, the committee's president and executive director, who added that educators and police officials nationwide should think carefully before entrusting children's safety to AI. Leah Frazier, director of the committee's Digital Justice Initiative, said the incident should never have been escalated to police at all.

What the plaintiffs are asking for

Beyond relief for the four students — Taki Allen, Yui-Osa Ogbeide, Devon Thompson Jr. and Keith Winfield III — the lawsuit seeks procedural reform. The committee wants police and schools required to verify reports from AI threat-detection platforms before acting, an evaluation of whether such technology belongs in schools at all, and policies ensuring that false or cancelled alerts cannot set off armed police responses. It is also calling for more training, oversight and transparency around errors and outcomes, and for community input into how these systems are deployed.

Why it matters

AI gun-detection and other surveillance products are spreading through US schools, typically marketed as proactive safety tools. This case documents what can happen when one gets it wrong: a confidence figure of 90 percent attached to a bag of chips ended with four teenagers on their knees in handcuffs.

Just as significant is the sociotechnical nature of the failure. The model made the initial error, but vendor staff, a school principal and a police department all missed or disregarded clear evidence that there was no weapon. Automated safety systems are only as safe as the human checkpoints built around them, and here those checkpoints appear to have been absent.

The committee also frames this as a civil rights issue, noting that inaccuracies and harms from AI in public safety fall disproportionately on Black communities and other people of color, especially in criminal justice settings. With districts across the country weighing similar tools, the outcome in Baltimore County could influence how verification and escalation rules are written elsewhere.

  • #ai-surveillance
  • #civil-rights
  • #education
  • #law-enforcement
  • #false-positives

Related posts