An incident at a high school in Baltimore, USA, has raised new concerns about the accuracy of AI surveillance tools used in educational institutions.
An incident at a high school in Baltimore, USA, has raised new concerns about the accuracy of AI surveillance tools used in educational institutions.
On Monday night, Taki Allen was sitting outside his school with friends after football practice, eating Doritos. About 20 minutes later, he saw eight police cars pull up, Interesting Engineering reports .
Of course, the guy couldn't even imagine that the police had come because of him until he was ordered to kneel, handcuffed, and searched.
"They made me kneel down, put my hands behind my back, and handcuffed me. Then they searched me and found that I had nothing on me," Allen recalls. "Then they went over to where I was standing and found a bag of chips on the floor."
When officers showed him the image that triggered the AI system's alarm, Allen was stunned: "I was just holding a bag of Doritos with two hands and one finger out, and they said it looked like a gun."
Last year, Baltimore County Public Schools began implementing Omnilert, an AI-powered gun detection system. The technology scans existing surveillance footage and automatically sends alerts to the school's security team and law enforcement if it detects an object it believes is a weapon.
After confirming that no weapon was present, the student was released. The principal said psychologists will provide direct support to Allen and other students who witnessed the incident.
But the incident has reignited the debate over the reliability of AI in real-world policing and school safety. What was supposed to prevent violence ended up turning an ordinary moment into a terrifying encounter with armed law enforcement.
Previously, dev.ua reported on an American schoolgirl who was visited by the police in the middle of the night due to an AI system error. The program decided that the girl wanted to harm herself.
