Back to News
SecurityAI Understanding briefing

Four Black teens sue over AI gun detector that mistook Doritos bag for a firearm

Four Black teenagers filed a lawsuit claiming Omnilert’s AI gun‑detection system falsely identified a bag of Doritos as a gun, prompting an armed police response at a Maryland high school.

4 min readRead the linked source
Source-page capture accompanying Four Black teens sue over AI gun detector that mistook Doritos bag for a firearm
Source referenceSource recorded
Publisher
cleveland.com
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Key terms

Algorithmic Bias
Systematic unfairness in model outputs caused by skewed data, assumptions, or modeling choices.
AI Safety
A field focused on reducing harmful behavior, failures, and misuse risks in AI systems.
Prompt
The input instructions and context provided to a generative model.
Test yourselfAI Ethics Quiz

What happened

Four Black teenagers were handcuffed and searched after Omnilert’s AI gun‑detection system issued an alert that a bag of Doritos was a firearm outside Kenwood High School in Baltimore County, Maryland. Police arrived with guns drawn, forced the students to their knees, and conducted a search before determining no weapon was present. The students have filed a lawsuit seeking damages and an injunction against the school district’s use of the system until better safeguards are put in place.

On Oct. 20, 2025, outside Kenwood High School in Baltimore County, Maryland, Omnilert’s AI gun‑detection system generated an alert indicating a 90% probability that a firearm was present. The system had analyzed an image of a bag of Doritos, which it mistakenly classified as a gun.

Police officers responded to the alert, arriving with weapons drawn. According to the complaint, officers forced the four Black teenagers—waiting for a ride after football practice—to kneel, handcuffed them, and conducted a full-body search before discovering no weapon.

School and police personnel later determined the alert was invalid and cancelled it, but the initial response had already occurred. The students filed a lawsuit alleging the AI system’s false alert led to an unlawful seizure and racial profiling.

Omnilert, the vendor of the AI system, stated that the image closely resembled a gun and that the system operated as designed. The company said its involvement ended once the incident was marked resolved, and it did not dispute the false‑positive nature of the alert.

The lawsuit seeks unspecified damages and asks a judge to halt the school district’s use of Omnilert’s system until the vendor implements better training, safeguards, and communication protocols to prevent similar incidents.

Source details: cleveland.com

Why it matters

The case highlights the real‑world consequences of false positives from AI‑driven security systems in schools, raising concerns about civil rights, racial bias, and the adequacy of current oversight. The lawsuit underscores the need for transparent validation, robust training data, and clear protocols for law‑enforcement response to AI alerts. If courts limit or reshape the deployment of such systems, it could affect how school districts nationwide adopt AI‑based safety tools, potentially prompting stricter regulatory scrutiny and prompting vendors to improve accuracy and accountability mechanisms.

The incident illustrates how AI‑driven security tools can produce high‑stakes false positives, especially in environments with vulnerable populations such as schools.

Racial dynamics are central: the plaintiffs are Black teenagers, and the case raises questions about whether AI systems disproportionately misidentify threats involving minority students, echoing broader concerns about .

Legal outcomes could set precedents for liability when AI vendors and public agencies rely on automated alerts that lead to police action, potentially influencing insurance, procurement, and risk‑management practices for school districts.

The case may policymakers to consider stricter standards for AI weapon‑detection accuracy, mandatory human verification steps, and clearer protocols for law‑enforcement response to AI alerts.

For vendors like Omnilert, the lawsuit could drive product redesigns, more rigorous testing, and transparency measures to maintain market confidence and avoid future litigation.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
Interactive Concept Check+10 Points
AI Ethics Quiz

Why can ethical evaluation not be reduced to one model score?

What to watch next

Future court rulings on this case, any policy changes by school districts or state education boards regarding AI surveillance, and Omnilert’s response or product revisions. Watch for legislative proposals on AI‑driven weapon detection in schools and any broader civil‑rights investigations into disparate impacts on minority students.

Court rulings on the plaintiffs’ claims, especially any findings on negligence or civil‑rights violations.

Legislative or regulatory actions at the state or federal level addressing AI‑based weapon detection in schools.

Potential settlements or policy changes announced by the Baltimore County school district or Omnilert.

Reactions from civil‑rights groups and education advocacy organizations, which may push for broader bans or stricter oversight of AI surveillance tools.

Industry response, including whether other vendors adjust their alert thresholds, verification processes, or public communications.

Related guides & quizzes

Found this useful?