An AI-powered weapon detection system at a Florida middle school falsely identified a student's clarinet as a rifle, triggering a lockdown and police response.
On a recent day at Lawton Chiles Middle School in Oviedo, Florida, an AI-powered surveillance system operated by ZeroEyes detected what appeared to be a person in camouflage holding a suspected weapon pointed down a hallway in a rifle position. This triggered an immediate alert to police, who responded and placed the school in lockdown. An officer searched classrooms but found no threat. Upon closer review, dispatchers informed the officer that the suspected rifle might have been a band instrument. The officer then located the source in the band room: a student wearing a military costume for a Christmas-themed dress-up day, holding a clarinet. ZeroEyes, the security company that operates the threat-detection system for Seminole County Public Schools, acknowledged the incident was caused by their AI system mistakenly flagging the clarinet as a weapon. The student was dressed as a military character from the movie 'Red One' and was unaware he was holding his instrument in a way that would trigger an alert. Police took no further action against the student. ZeroEyes co-founder Sam Alaimo defended the alert, stating the student appeared to be aiming the clarinet like a gun and that it was better to dispatch police than not. The incident is part of a broader trend of AI threat-detection systems in schools producing false alarms, including similar incidents in Maryland and Texas.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed