AI Weapon Scanners in NYC Subway Yield False Positives

The Evolv AI scanning pilot was actively in use and its malfunction led to real-world harm—false alarms causing invasive searches and privacy violations—making this a concrete AI Incident under the framework.

The weapons scanner is an AI system whose malfunction (high false positive rate) directly caused harm—unwarranted invasive searches of 118 passengers—constituting a breach of human rights (privacy) under the AI Incident definition.

The AI system (weapon detection) was explicitly used and malfunctioned, producing false positives and failing to detect concealed weapons, which led to unnecessary searches and potential violations of constitutional rights (Fourth Amendment). This constitutes a violation of human rights and a breach of legal protections, fitting the definition of an AI Incident. The investigations into the company's misconduct and the city's handling of the contract further contextualize the incident but do not change the classification. The harm is realized, not just potential, so this is not merely a hazard or complementary information.

The article explicitly involves an AI system (Evolv's AI-powered gun detection scanners) used in a public safety context. The system's deployment and malfunction (high false positive rate and failure to detect guns) directly led to ineffective security screening, which can harm community safety and trust, a form of harm to communities. The CEO's admission that the technology is not suited for subway environments further confirms the malfunction aspect. The event meets the criteria for an AI Incident as the AI system's use has directly led to harm (ineffective security and potential public safety risks).

The AI system (Evolv scanners) was actively used, but the pilot resulted only in false positives without any detected firearms or harm. The article discusses the system's reliability issues, public debate, and legal scrutiny, which are responses and contextual information rather than direct or indirect harm caused by the AI system. No injury, rights violation, or other harms occurred, and no plausible future harm is clearly indicated. Hence, the event does not meet the criteria for AI Incident or AI Hazard but fits the definition of Complementary Information.

The AI system (weapons scanners using AI) is explicitly mentioned and is being used in a real-world setting. The event involves the use of AI systems for security screening, which could plausibly lead to harm such as wrongful police interactions or rights violations due to false positives. However, the article does not report any actual injury, rights violation, or other harm occurring so far. The concerns and criticisms are about potential harms and effectiveness, making this a plausible risk rather than a realized incident. Therefore, this qualifies as an AI Hazard rather than an AI Incident. It is not merely complementary information because the main focus is on the potential risks and criticisms of the AI system's deployment, not on responses or ecosystem context. It is not unrelated because AI systems are central to the event.

The AI system (weapons detection scanners using AI) is explicitly mentioned and is in use. The event involves the use of the AI system in a real-world setting (transit system). While there are many false positives that could plausibly lead to harm (e.g., confrontations with police, rights violations), the article does not report any actual injuries, rights violations, or other harms occurring so far. The concerns and criticisms suggest potential future harms, making this an AI Hazard rather than an AI Incident. The event is not merely complementary information because it focuses on the pilot's outcomes and potential risks, not just updates or responses to past incidents.

The AI-powered weapons scanners are clearly AI systems involved in the event. The event concerns their use and performance (including false positives) but does not report any direct or indirect harm resulting from their deployment. There is no indication that the AI system caused injury, rights violations, or other harms, nor that it plausibly could have led to such harms in this pilot. The event mainly provides results of the pilot test and public/legal reactions, which fits the definition of Complementary Information. Hence, the classification is Complimentary Info.