This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A 61-year-old man was wrongfully arrested for armed robbery after a retail loss prevention agent used facial recognition software to misidentify him. While held in jail for ten days, the man was beaten and sexually assaulted by other inmates. He is now suing the retail companies involved for $10 million, highlighting the dangers of relying on error-prone facial recognition technology for law enforcement investigations.
Harvey Murphy Jr. was wrongfully accused of robbing a Sunglass Hut due to an alleged misidentification by the facial recognition system operated by Macy's. While in custody for ten days, he was sexually assaulted. He is now suing Macy's, EssilorLuxottica (Sunglass Hut's parent), and others, for $10 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.
Also classified as
The reports highlight that facial recognition technology has systemic biases and unequal accuracy rates across different demographic groups.
Loss prevention agents and police officers reflexively trusted and over-relied on the automated facial recognition match without sufficient independent verification.
The facial recognition system failed to perform reliably under varying conditions, generating a false match from low-quality video that led directly to a wrongful arrest.
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
The incident was initiated by a faulty match produced by a deployed facial recognition AI system, which was an unexpected and erroneous outcome.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
A domestic incident where an erroneous facial recognition match by private retailers led to the wrongful arrest and subsequent severe abuse of an individual in jail. While highlighting serious civil rights and systemic AI reliability concerns, the incident has negligible direct national security implications.
The facial recognition software operates as a tool. It performs a specific, well-defined function of matching faces from surveillance footage against a database based on direct input, without autonomous planning or context retention.
High Risk: The system is a law enforcement application, specifically facial recognition used for identifying suspects, which is explicitly listed as a High Risk application under the EU AI Act.
The system's primary function was to scan surveillance footage and match faces against an internal database to identify individuals suspected of retail theft.
The reports provide consistent, detailed accounts of the incident, including the victim's name, age, location, the retail stores involved, the nature of the AI failure, and the specific physical and psychological harms suffered. The legal filings and statements from attorneys provide strong factual grounding.
None. The causal chain from the faulty AI match to the wrongful arrest and subsequent assault is clear and undisputed by the available reports.
Technology
Unreliable Facial Recognition
The facial recognition software is highly error-prone and unreliable.
Sensitivity to Image Quality
The software's accuracy is highly dependent on camera quality and lighting.
Data Inputs
Low-Quality Surveillance Video
The input video came from low-quality cameras, leading to false matches.
Outdated Database Mugshots
The database scanned old mugshots from the 1980s, causing misidentification.
Human Factors
Automation Bias
Police reflexively trusted automated facial recognition results without doubt.
Witness Priming
Confidence in the AI results primed the witness to make a false identification.
Process and Methods
Prematurely Halted Investigation
Loss prevention told police to stop investigating after getting an AI match.
Flawed Photo Lineup Process
The police conducted a photo lineup based solely on the faulty AI lead.
Management
Reliance on Faulty Technology
Management deployed facial recognition despite knowing it was error-prone.
Lack of Safeguards
No corporate safeguards to verify AI matches before reporting to police.
Regulatory Environment
Lack of Retail AI Regulation
Absence of strict regulations allowed unchecked retail use of risky FRT.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
