Walmart's Bagging-Detection False Positives Exposed Workers to Health Risk

Walmart's theft-deterring bagging-detection system allegedly exposed workers to health risks during the coronavirus pandemic when its false positives prompted workers to unnecessarily step in to resolve the issue.

Risk classification

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

The incident was primarily caused by the Everseen AI system's lack of robustness, specifically its high rate of false positives that incorrectly flagged legitimate transactions.

Causal factors

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

The risk arose from the AI system's unintentional false positive errors occurring after its deployment in retail stores.

EU AI Act risk tier

• Risk tier: 4 Minimal or No Risk

Minimal or No Risk: The system is a retail theft-deterrence tool, which generally poses low risk to users and society under normal conditions, falling outside the high-risk categories defined in the Act.

AI system and alleged parties

• AI system: Everseen AI platform (Everseen)
• AI purpose: Threat Detection; Camera Tracking
• Behaviour type: Autonomous
• Alleged developer: Everseen
• Alleged deployer: Walmart
• Alleged harmed parties: Walmart employees

Harm severity

Highest direct severity in any category: Severe. 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 Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Psychological

Reported: The report explicitly describes that the system became a serious concern and a source of anxiety for employees during the pandemic.

Directly caused: N/A

Indirectly caused: Employees experienced heightened stress and anxiety over their personal safety and health because the faulty system forced them into close proximity with customers.

Inferred additional harm: Widespread distress and feelings of unsafety likely affected self-checkout hosts across the 2,000+ stores where the system was active.

People affected

• Occurrences reported: 1
• People reportedly exposed: 2000

Potential causes

Management

• Prioritizing Asset Protection: Management prioritized inventory loss prevention over employee health.
• Failure to Suspend System: Corporate chose to keep the error-prone system active during the pandemic.

Technology

• Frequent False Positives: The system flagged legitimate scans as 'non-scans' triggering alerts.
• Computer Vision Limitations: Algorithms struggled to accurately differentiate bagging from scanning motions.

Data Inputs

• Unsynchronized Motion Data: Discrepancies between camera tracking and register logs caused false alerts.

Human Factors

• Forced Physical Proximity: Alerts forced employees to approach customers, violating social distancing.
• Varied Shopper Behaviors: Diverse customer handling of items led to misinterpretation by the AI.

Process and Methods

• Removal of Weight Sensors: Replacing physical sensors with AI removed a backup validation method.
• Proximity-Based Alert Resolution: The resolution protocol required physical intervention rather than remote checks.

Regulatory Environment

• Inadequate Pandemic Mandates: Lack of enforceable retail safety rules allowed high-contact systems to remain.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, detailed accounts of the AI system's deployment, its specific technical failure (false positives), and the resulting real-world safety risks for employees. The causal link between the AI's errors and the increased health risk is well-established.
• Ambiguities identified: The exact rate of false positives is not quantified, and there is no direct data linking specific COVID-19 infections or deaths among Walmart employees to the Everseen system's false alarms.
• Alternative interpretations: The incident could be viewed primarily as a labor management issue regarding pandemic safety protocols rather than an AI safety failure, though the AI's technical inaccuracy was the direct trigger for the safety protocol violations.

Walmart's deployment of Everseen's AI theft-deterrence system during the COVID-19 pandemic resulted in high false-positive rates, forcing employees into close contact with customers. While this raised valid workplace safety and public health concerns for essential workers, the incident remains a commercial operational failure with negligible national security implications.

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

Threat characteristics

• Imminence: Long-term. The event represents a historical operational and labor safety issue rather than an active, imminent national security crisis.
• Autonomy: Human-supervised. The computer vision system autonomously flagged potential theft incidents, but required human employees to intervene and resolve the alerts.
• Novelty: Established threat. False positives in commercial computer vision and automated checkout systems are well-known and established technical limitations.

Impact by dimension

• Physical security: Negligible. The incident involved a retail theft-detection system causing operational friction and social distancing issues for grocery workers. It posed no threat to physical critical infrastructure or national security.
• Information security: Negligible. No intelligence compromise, classified data theft, or information warfare operations were associated with this commercial retail incident.
• Sovereignty: Negligible. The incident was entirely confined to commercial retail operations and did not impact state authority, electoral systems, or government functions.
• Economic security: Negligible. While involving retail inventory management, the incident did not threaten national financial systems, strategic technology, or economic security.
• Societal stability: Negligible. Although employees raised concerns about workplace safety during the pandemic, the incident did not represent a systemic threat to societal stability, civil liberties, or human rights.