HireVue and the EU AI Act: full analysis
Why HireVue is classified as high-risk
The EU AI Act's Annex III, Category 4 covers AI systems "intended to be used for recruitment or selection of natural persons, in particular to analyse and filter job applications, and to evaluate candidates in the course of interviews or tests." HireVue's core product — AI-powered video interview scoring that evaluates verbal responses, communication patterns, and structured interview performance — sits squarely in this category.
The classification applies regardless of how HireVue markets its product. What matters is function: if the AI system evaluates a candidate and that evaluation influences hiring decisions, it is high-risk.
Provider vs deployer: who is responsible for what?
| Obligation | HireVue (provider) | Your company (deployer) |
|---|---|---|
| Technical documentation (Annex IV) | ✓ Their responsibility | Verify it exists |
| Conformity assessment | ✓ Their responsibility | Request Declaration of Conformity |
| EU database registration | ✓ Their responsibility | Confirm it is registered |
| Candidate disclosure notice | — | ✓ Your responsibility |
| Human oversight procedures | — | ✓ Your responsibility |
| Log retention (6 months) | — | ✓ Your responsibility |
| Worker notification | — | ✓ Your responsibility |
| GDPR DPIA for candidate data | — | ✓ Your responsibility |
What "meaningful human oversight" looks like in practice
Article 26(2) requires that the person making the hiring decision has genuinely reviewed the AI output, understands what the system assessed, and exercises independent judgement. It does not mean a human simply signs off on an AI-generated shortlist without reviewing the underlying evidence.
In practice, this means:
- Recruiters must understand what HireVue scored and how, not just the overall rank
- A human must be able to override any AI recommendation
- The override (or acceptance) must be documented
- Recruiters may not reject candidates based solely on HireVue scores without human review
Candidate disclosure: what to say and when
You must inform candidates before or during the interview process that an AI system will be used to evaluate their performance. The disclosure cannot be buried in terms of service. A compliant approach:
- Include a clear statement in the interview invitation email: "Your interview will be evaluated using AI-powered assessment tools. A hiring manager will review all AI assessments before any decision is made."
- Display a disclosure at the start of the HireVue session
- Provide a contact point for candidates to request an explanation of the AI assessment
The deadline situation
High-risk obligations officially apply from 2 August 2026. A provisional agreement reached on 27 May 2026 (Digital Omnibus) proposes deferring this to 2 December 2027 — but this has not been formally adopted into law. Legal advisors recommend continuing preparation against the 2 August 2026 deadline. The candidate disclosure and worker notification obligations should be implemented regardless.