Does AI interview proctoring have to comply with NYC's Local Law 144 or the EU AI Act?
TL;DR: It depends on what the tool actually does. Laws like NYC Local Law 144 and the EU AI Act target AI that makes or substantially assists a hiring decision. A proctoring tool that flags integrity concerns for a human to review — without scoring or ranking candidates — sits in a different, lower-risk category than an AI resume screener or automated interview scorer.
The Evidence
NYC Local Law 144, in effect since July 2023, regulates "Automated Employment Decision Tools" (AEDTs) — systems that substantially assist or replace discretionary decision-making in hiring or promotion. Covered tools require an annual independent bias audit, published results, and advance notice to candidates. The law's own guidance draws a distinction between AI that scores or ranks candidates and AI that simply provides information a human then evaluates.
The EU AI Act, which entered into force in August 2024 with obligations phasing in through 2026–2027, classifies AI systems used "to make decisions on or materially influence" recruitment and selection — including targeted job ads, screening, and evaluation — as high-risk, triggering conformity assessments, transparency requirements, and mandatory human oversight. Employment-related AI is one of the Act's explicitly named high-risk categories.
Where a given proctoring tool lands depends on its function, not its label. A system that assigns a hire/no-hire recommendation based on interview behavior would likely be in scope for both frameworks. A system that flags a violation timeline and evidence — coaching detected, identity mismatch, second voice — for a human reviewer to weigh alongside everything else they know about the candidate is providing information, not making or substantially assisting the decision itself.
Where Different Tool Types Land
| Tool behavior | NYC LL144 | EU AI Act |
|---|---|---|
| Ranks/scores candidates automatically | Likely covered (AEDT) | Likely high-risk |
| Rejects candidates without human review | Likely covered | Likely high-risk |
| Flags integrity evidence for human review only | Lower risk, evaluate case by case | Lower risk, evaluate case by case |
| No candidate evaluation, purely session monitoring | Generally outside AEDT scope | Generally lower obligation tier |
How to Stay Compliant
- Confirm whether your tool produces a hiring recommendation/score, or purely evidence for human review — this is the pivotal distinction under both frameworks.
- Give candidates advance notice that AI-based monitoring is part of the interview process, in writing.
- Keep a human decision-maker explicitly in the loop for any adverse outcome tied to a flagged session.
- Ask your vendor for their bias-audit documentation and data-handling practices in writing.
- Consult employment counsel for your specific jurisdiction — this article summarizes public regulatory frameworks, not legal advice for your situation.
FAQ
Is Neuroxa's AI Meeting Proctor an AEDT under NYC law? Neuroxa flags evidence — coaching, second voices, identity mismatches — for a human hiring team to review; it doesn't score, rank, or auto-reject candidates, which is the key distinction NYC's guidance draws. Confirm your specific use case with counsel.
Does the EU AI Act ban AI interview monitoring outright? No — it classifies certain employment-related AI as high-risk, requiring compliance obligations like transparency and human oversight, not a ban.
Do these laws apply outside NYC and the EU? Similar AI-hiring regulations are emerging in other US states and countries; check your specific jurisdiction.
What should I ask a proctoring vendor about compliance? Whether the tool makes hiring decisions or only surfaces evidence, what bias-audit documentation exists, and how candidate notice is handled.
By Pinal Dave Last updated: 2026-07-25