Three Make-Up Artists Lost Jobs Following Black-Box Automated Decision by HireVue

Three make-up artists lost their positions following an algorithmically-assessed video interview by HireVue who reportedly failed to provide adequate explanation of the findings.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.4 Lack of transparency or interpretability

The primary issue was the complete lack of transparency and explainability regarding how the HireVue algorithm scored the candidates, leaving them and managers unable to understand the redundancy decisions.

Additional risk subdomains

• 6.2 Increased inequality and decline in employment quality: The use of automated AI assessments to make redundancy decisions directly impacted the employment status and job security of the workers.
• 7.3 Lack of capability or robustness: The visual analysis component was discontinued because it provided marginal value, indicating a lack of robust capability for job performance evaluation.

Causal factors

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

The incident arose from the deployment of the HireVue AI system, which produced unexpected and unexplainable redundancy recommendations that led to the dismissal of qualified employees.

EU AI Act risk tier

• Risk tier: 2 High Risk

High Risk: The AI system was used for employment and worker management (recruitment and redundancy assessment), which is explicitly classified as High Risk under the EU AI Act.

AI system and alleged parties

• AI system: HireVue
• AI purpose: Application Evaluation; Workforce Monitoring and Evaluation
• Behaviour type: Assistant
• Alleged developer: HireVue
• Alleged deployer: Estée Lauder
• Alleged harmed parties: pseudonymous Estée Lauder's former staff

Harm severity

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

Financial

Reported: The report describes 1 incident causing financial loss to three employees who lost their jobs, followed by an out-of-court settlement of undisclosed value.

Directly caused: The three make-up artists suffered immediate loss of employment and regular income due to the automated redundancy decision.

Indirectly caused: The employees likely incurred legal costs to challenge the redundancy, which were later resolved through an out-of-court settlement.

Inferred additional harm: N/A

Differential treatment

Reported: The report explicitly describes differential treatment where employees were evaluated and made redundant based on algorithmic video analysis.

Directly caused: Three make-up artists were selected for redundancy based on an automated HireVue assessment that analyzed their expressions and verbal answers rather than practical demonstrations.

Indirectly caused: N/A

Inferred additional harm: Other employees undergoing the same redundancy process were also subjected to this automated evaluation, potentially leading to wider unfair treatment.

Psychological

Reported: The report explicitly describes psychological distress and self-doubt experienced by the affected employees.

Directly caused: One employee, PERSON_001, reported doubting herself massively and finding it really hard to be told she was not good enough by an algorithm.

Indirectly caused: N/A

Inferred additional harm: It is highly likely that the other two dismissed employees, Anthea and Onieka, experienced similar feelings of distress, anxiety, and self-doubt due to their unexplained job losses.

People affected

• Occurrences reported: 1
• People reportedly harmed: 3
• People reportedly exposed: 3

Potential causes

Management

• Inappropriate Tool Selection: Management selected a generic video AI tool for highly practical roles.
• Defensive Corporate Response: Management deflected blame and claimed the AI played a non-decisive role.
• Lack of Algorithmic Transparency: Failure to explain to employees how they were being evaluated by the AI.

Technology

• Flawed Algorithmic Assessment: Algorithm evaluated speech and expressions without practical demonstration.
• Low Value Visual Analysis: HireVue used facial analysis which was later admitted to have minimal value.
• Opaque Decision-Making Logic: System used 15000 complex data points making decisions unexplainable.

Data Inputs

• Lack of Practical Performance Data: System relied on verbal descriptions instead of physical makeup demonstrations.
• Unreliable Facial Expression Data: Facial recognition data was used despite having marginal predictive utility.

Human Factors

• Overreliance on Automation: Managers trusted automated scores without manually reviewing video interviews.
• Difficulty with Verbal Assessment: Candidates struggled to verbally describe physical tasks like applying makeup.
• Negative Psychological Impact: Rejection by an opaque algorithm caused candidates to deeply doubt themselves.

Process and Methods

• Inadequate Appeal Feedback: Appeals received generic copy-pasted explanations about AI algorithms.
• No Manual Video Review: No humans watched the interview videos to verify algorithmic scoring.
• Inappropriate Evaluation Format: Using a structured video QandA to assess hands-on artistic skills.

Regulatory Environment

• Lack of Algorithmic Auditing Rules: No regulatory requirements to audit hiring AI for fairness or validity.
• Absence of Explainability Mandates: No legal framework forced companies to provide clear algorithmic explanations.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent details about the three employees, the AI vendor (HireVue), the employer (Estee Lauder/MAC), the outcome of the automated assessment, and the subsequent legal settlement. The role of the AI in the redundancy process is explicitly documented.
• Ambiguities identified: There is some conflict regarding the exact weight of the facial recognition component, with Estee Lauder claiming it was under 1% while the employees felt it was a decisive factor.
• Alternative interpretations: None. The incident is clearly an employment dispute arising from algorithmic decision-making.

Three MAC makeup artists in the UK lost their jobs due to an opaque automated video interview algorithm by HireVue. The employment dispute was settled out of court, and HireVue subsequently disabled its visual analysis feature. The incident carries negligible national security implications, representing a localized corporate labor issue rather than a threat to national infrastructure, sovereignty, or security.

• Overall national security impact: Negligible
• Response level: Minor
• Scope: Single nation
• Primary target: United Kingdom
• Alleged perpetrator: Estee Lauder

Threat characteristics

• Imminence: Long-term. The incident is a resolved historical event with no active or imminent threat to national security.
• Autonomy: Human-supervised. The AI system provided automated scores used in tandem with human decision-making by MAC management.
• Novelty: Evolved capability. Represents an advancement in the deployment of automated facial and verbal analysis for employment decisions, leading to subsequent feature removal.

Impact by dimension

• Physical security: Negligible. No physical systems, critical infrastructure, or human safety were threatened or impacted by this commercial HR software.
• Information security: Negligible. The incident involved corporate HR recruitment software and did not impact intelligence capabilities or involve information warfare.
• Sovereignty: Negligible. Private sector employment dispute with no impact on state authority, government decision-making, or constitutional processes.
• Economic security: Negligible. While three individuals suffered personal financial loss, there was no threat to strategic industries or national technological security.
• Societal stability: Negligible. The incident affected only three individuals and was resolved through an out-of-court settlement, presenting negligible threat to national societal stability.