AI Gun Detection System Mistook Doritos Bag for Weapon, Student Detained
An AI gun detection system deployed at a Baltimore high school mistakenly identified a 16-year-old student's empty snack bag as a firearm, resulting in the student being detained at gunpoint and handcuffed by police. The school superintendent stated the system functioned as designed by alerting humans for verification, though the incident demonstrates how such alerts can trigger armed responses before confirmation occurs.
What happened
On October 2025, an AI gun detection system at Kenwood High School in Baltimore, Maryland mistook a 16-year-old student's empty Doritos bag for a firearm. Taki Allen had crumpled the bag after eating the chips and placed it in his pocket while waiting for pickup after football practice.
Police arrived at the school with approximately eight cars and officers approached Allen with guns drawn, ordering him to get on the ground. Officers handcuffed and searched Allen, eventually finding the crumpled snack bag, and an officer explained that the school's AI detection system had flagged the bag as appearing to be a weapon.
Reported impact
Allen stated that the experience was frightening and that he was made to get on his knees, handcuffed, and searched. He reported feeling uneasy at school following the incident and said he avoided going outside after practice.
Allen said the school principal did not speak to him until three days after the event, which he felt suggested the school did not genuinely care about his wellbeing.
Index analysis
Baltimore County Public Schools has deployed the Omnilert AI gun detection system since 2023, which uses existing school cameras to identify people and objects. The system automatically sends alerts to school officials and police if a weapon is detected.
School superintendent Dr Myriam Rogers stated the system functioned as intended by alerting humans for verification rather than taking autonomous action. Rogers said the program is 'based on human verification' and 'did what it was supposed to do, which was to signal an alert for humans to take a look.'
Governance controls
The school's superintendent defended the system's design, emphasizing that human verification is built into the process. However, the incident demonstrates that alerts generated by the system, even when based on false identifications, can trigger armed police responses before verification occurs.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.