Alleged Macy's Facial Recognition Error Leads to Wrongful Arrest and Subsequent Sexual Assault in Jail

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.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

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.

Additional risk subdomains

• 5.1 Overreliance and unsafe use: Loss prevention agents and police officers reflexively trusted and over-relied on the automated facial recognition match without sufficient independent verification.
• 1.3 Unequal performance across groups: The reports highlight that facial recognition technology has systemic biases and unequal accuracy rates across different demographic groups.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The incident was initiated by a faulty match produced by a deployed facial recognition AI system, which was an unexpected and erroneous outcome.

EU AI Act risk tier

• Risk tier: 2 High Risk

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.

AI system and alleged parties

• AI system: facial recognition system
• AI purpose: Face Recognition; Identification
• Behaviour type: Tool
• Alleged developer: unknown, Macy's
• Alleged deployer: Macy's
• Alleged harmed parties: Harvey Murphy Jr

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Minor
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Minor
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Minor, indirect Minor
• Civil rights: direct Negligible, indirect Minor
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Minor
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Negligible, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes severe physical harm to the victim resulting from his wrongful detention.

Indirectly caused: The victim was brutally beaten and gang-raped by three men in a county jail, leaving him with permanent and lifelong physical injuries.

Financial

Reported: The report notes that the victim has filed a lawsuit seeking 10 million dollars in damages.

Indirectly caused: The victim likely incurred legal fees and lost income during his 10-day incarceration, and is seeking 10 million dollars in pending litigation.

Differential treatment

Reported: The report explicitly describes differential treatment of individuals based on AI outputs.

Directly caused: The AI system singled out the victim based on a faulty match, leading to his targeted arrest and detention.

Indirectly caused: The reports discuss how facial recognition technology disproportionately misidentifies and impacts minority groups, particularly Black individuals.

Civil rights

Reported: The report explicitly describes severe violations of human and civil rights.

Indirectly caused: The victim's fundamental civil rights were violated through wrongful arrest, deprivation of liberty, and detention without bond for 10 days based on a faulty AI match.

Privacy

Reported: The report explicitly mentions privacy concerns and previous lawsuits regarding biometric data collection.

Indirectly caused: The retail stores scanned surveillance footage against databases containing old mugshots and biometric data without explicit consent, violating biometric privacy norms.

Psychological

Reported: The report explicitly describes severe psychological and mental-health harm to the victim.

Indirectly caused: The victim suffered extreme anxiety, terror, and ongoing trauma, describing it as a 'day-to-day personal pain in your heart and your soul' and feeling 'terrified by the idea that it could happen again.'

Epistemic

Reported: The report explicitly describes epistemic harm where a false AI output was treated as factual truth.

Directly caused: The AI system generated a false match, creating a false record that the victim committed armed robbery.

Indirectly caused: The false match misled loss prevention agents, police investigators, and witnesses, priming them to believe the victim was guilty without substantial evidence.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1

Potential causes

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.

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.

Regulatory Environment

• Lack of Retail AI Regulation: Absence of strict regulations allowed unchecked retail use of risky FRT.

Information quality

• Classification confidence: High
• Reason for confidence: 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.
• Ambiguities identified: The exact developer or brand of the facial recognition software used by the retailers is not specified in the reports.
• Alternative interpretations: None. The causal chain from the faulty AI match to the wrongful arrest and subsequent assault is clear and undisputed by the available reports.

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.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The active crisis of the wrongful arrest is resolved, but the systemic risk of facial recognition errors remains an ongoing long-term concern.
• Autonomy: Human-controlled. The facial recognition system acted as an investigative tool, with human officers and retail agents retaining final decision-making authority over the arrest.
• Novelty: Established threat. Facial recognition misidentification leading to wrongful arrest is an established threat with several previously documented cases.

Impact by dimension

• Physical security: Negligible. No threats to physical systems, critical infrastructure, or kinetic military capabilities were identified in this domestic law enforcement incident.
• Information security: Negligible. No information warfare, intelligence compromise, or systematic state-sponsored disinformation campaigns were involved.
• Sovereignty: Minor. Minor impact on local judicial and law enforcement functions due to reliance on private sector AI tools, but no threat to state authority or national sovereignty.
• Economic security: Negligible. No threat to strategic industries, financial systems, or national technological competitive advantage; impact is limited to private civil litigation.
• Societal stability: Minor. Represents a severe localized violation of individual civil and human rights due to AI misidentification, but does not threaten large-scale national societal stability.