Evolv Technology's AI-powered weapons detection scanners deployed in Charlotte Mecklenburg School District failed to reliably detect weapons while creating operational chaos through excessive false positives and requiring significantly more security personnel than promised.
In early 2022, Charlotte Mecklenburg School District purchased Evolv Technology's AI-based weapons screening system for $4.9 million to address gun incidents across its 21 high schools. The system promised 'line-free' scanning that would detect concealed weapons using artificial intelligence while allowing students to pass through without stopping or emptying pockets. However, internal emails from nearly 2,000 documents obtained through public records requests revealed significant problems during deployment. The scanners failed to detect a Glock pistol at commonly-used sensitivity settings, while simultaneously triggering false alarms on everyday school items like 3-ring binders, laptops, and notebooks. One principal described the first day as 'probably the least safe day' because the machines required manual searches of 'almost every child' and monopolized security staff attention at entrances instead of monitoring hallways. During the 2021-22 school year, only one gun was detected by Evolv scanners compared to 30 guns found by other means. The district had to hire 53 additional security guards to manage the increased workload, contradicting Evolv's promise of requiring fewer staff. Despite these failures, the district expanded the program to middle schools for an additional $9.9 million.
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