HireVue dropped facial-expression analysis after EPIC and the ACLU raised AI bias concerns
HireVue discontinued the facial expression analysis component of its AI video interview screening tool in January 2021 after EPIC filed an FTC complaint alleging unfair and deceptive practices, and senators Elizabeth Warren and Bernie Sanders raised bias concerns. The system analyzed facial microexpressions to score candidates on traits like emotional intelligence and dependability, but critics warned it systematically disadvantaged people with disabilities such as autism and Bell's Palsy and produced higher error rates for people of color. HireVue retained speech and language analysis but acknowledged the facial component was not worth the concern it generated.
Facial expression algorithms cannot reliably infer psychological traits from faces, systematically penalizing candidates whose expressions deviate from training data norms due to disability or race.
Key facts
- What
- HireVue discontinued the facial expression analysis component of its AI video interview screening tool in January 2021 after EPIC filed an FTC complaint alleging unfair and deceptive practices, and senators Elizabeth Warren and Bernie Sanders raised bias concerns.
- Incident date
- Jan 12, 2021
- Who
- HireVue
- Failure mode
- Policy Violation
- AI surface
- Agentic Workflow
- Severity
- High
What happened
HireVue deployed an AI-powered video interview platform that analyzed facial expressions, speech intonation, and word choice to score job candidates on traits such as emotional intelligence, cognitive ability, and dependability. EPIC filed an FTC complaint in October 2019 alleging unfair and deceptive practices, the ACLU of Massachusetts criticized the system as a discriminatory black box, and senators Elizabeth Warren and Bernie Sanders raised concerns about bias against people with disabilities and people of color. In January 2021, HireVue announced it would remove the facial expression analysis component while continuing to use speech and language analysis. HireVue's CEO acknowledged the facial analysis was adding some value for customers but was not worth the concern it generated.
What broke inside the model
- 01 · TriggerA prompt pushes against a deployment boundary.
- 02 · Model stepThe model produces the disallowed output.
- 03 · Control gapNo enforcement blocks it at generation time.
- 04 · FailureThe output crosses the policy line.
- 05 · ConsequenceA limit the business set is breached in public.
The output crosses a policy boundary the deployment had defined.
The facial expression analysis algorithm attempted to infer psychological traits and emotional states from facial microexpressions, a scientifically disputed methodology that cannot reliably map facial movements to internal states across different cultures, contexts, and conditions. Training data drawn from past hiring decisions perpetuated existing biases, and the system could not account for facial expressions affected by disabilities like Bell's Palsy, autism, or speech impediments, systematically devaluing those candidates. The algorithm also exhibited higher error rates for people of color due to unrepresentative training datasets.
What it cost
Sources
Cite this entry
https://failureindex.ai/failures/hirevue-dropped-facial-expression-analysisAI Failure Index. "HireVue dropped facial-expression analysis after EPIC and the ACLU raised AI bias concerns" (FI-0149). Realm Labs. https://failureindex.ai/failures/hirevue-dropped-facial-expression-analysis (indexed Jun 4, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0149. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
Realm compares what the model is about to output or do against the policy that governs the deployment, in real time, and can deny or redact the action before it takes effect, which is the gap an after-the-fact review never closes in time.