Baltimore student handcuffed after AI system mistakes bag of chips as weapon
Embedded Files
Occurred: October 2025
Page published: October 2025
Report incident๐ฅ| Improve page ๐| Access database ๐ข
A Baltimore County high school student was handcuffed after an AI surveillance system mistakenly identified his crumpled bag of chips as a gun, raising questions about the accuracy and reliability of the system.
What happened
What happened
Student Taki Allen was sitting outside Kenwood High School after football practice and was flagged by an AI-powered gun detection system that erroneously identified his Doritos bag of chips as a weapon.
The false alarm triggered a heavy police response, with approximately eight police cars arriving, officers drawing their guns, detaining, handcuffing, and searching the student, before realising the error.
The incident caused the student emotional distress, humiliation, and trauma due to false imprisonment and unnecessary physical restraint.ย
Why it happened
Why it happened
The alert was generated by AI weapons detection system Gun Detect, which flagged the bag of chips as dangerous and escalated the event for human review.
Developed by Omnilert, Gun Detect is designed to detect weapons using advanced visual recognition, but it misinterpreted the shape and positioning of the chip bag in Allen's hands.ย
The error highlights the lack of accuracy of the system and the lack of adequate human oversight in interpreting such critical alerts.ย
What it means
What it means
The incident reflects the dangers of overreliance on imperfect AI technology in sensitive environments like schools, and the failure to immediately recognise the alert as a false positive.
For the victim, it potentially means long-term psychological harm and loss of trust in school safety measures.
More broadly, it stresses the need for enhanced safeguards, robust human review mechanisms to prevent harms caused directly or indirectly by AI.
System ๐ค
System ๐ค
Gun Detect ๐
Developer: Omnilert
Country: USA
Sector: Education
Purpose: Detect weapons
Technology: Computer vision; Context recognition; Machine learning
Issue: Accuracy/reliability; Autonomy; Human/civil rights
AIAAIC Repository ID: AIAAIC2077
Page updated
Google Sites
Report abuse