AI-assisted hiring interviews are permitted in most jurisdictions, but several laws require candidate notice, bias audits, and transparent scoring. Here is what employers need to know.
Yes — AI video interviews are broadly legal in most countries, but legality depends on how the tool is used. Key obligations vary by location: some jurisdictions require prior notice to candidates, opt-out rights, bias audits, or data-processing disclosures. Employers should review applicable local law before deploying any AI screening tool. Criba's explainable, evidence-linked scoring is designed to support transparency and auditability wherever you hire.
Hiring technology has moved faster than legislation, but regulators are catching up. New York City's Local Law 144 requires annual bias audits for automated employment decision tools. Illinois's AIVEA mandates disclosure before AI is used to evaluate candidates. The EU AI Act classifies AI recruitment tools as high-risk, triggering transparency and human-oversight requirements. Other jurisdictions — Canada, California, and several others — have pending or recently enacted rules. The common thread: candidates should know when AI is evaluating them, employers should be able to explain results, and scoring systems should be auditable for bias. Criba is built around exactly these principles.
Most emerging regulations require informing candidates before AI evaluates them. Criba's candidate-facing interface makes the AI-assisted nature of the screening explicit, supporting pre-screening disclosure requirements.
Black-box AI scores are difficult to defend if challenged. Criba ties every Pass / Borderline / Reject outcome directly to candidate quotes, so recruiters can explain and document the basis for any decision.
Inconsistent human interviewing can introduce bias at scale. Criba applies the same structured criteria to every candidate, reducing the variance that often underlies disparate-impact claims.
NYC Local Law 144, the EU AI Act, and similar frameworks require auditable records. Criba's evidence-linked outputs give HR teams a documented, reviewable record of how screening decisions were reached.
Criba produces a ranked shortlist — it does not make the hire. Keeping a human in the loop satisfies the oversight requirements found in the EU AI Act and similar high-risk AI regulations.
Criba's ~5-minute voice-first interview collects only the information needed to rank candidates against role criteria, supporting data-minimization principles under GDPR and equivalent frameworks.
Review local requirements before you launch
Check whether your jurisdiction has specific rules — bias-audit obligations, disclosure timelines, opt-out rights. Laws in NYC, Illinois, the EU, and elsewhere differ in scope. When in doubt, consult employment counsel.
Disclose AI use to candidates upfront
Include a clear statement in your job posting or application flow that an AI tool will be used in screening. Criba's candidate-facing interface already surfaces this context; pair it with a line in your job ad.
Configure role criteria transparently
Define what competencies or skills the AI will evaluate before screening begins. Documented, role-specific criteria make it easier to demonstrate that your tool is measuring job-relevant attributes, not proxies for protected characteristics.
Keep humans accountable for final decisions
Use Criba's shortlist as an input to your recruiter review — never as an automatic gate. Document that a human reviewed the evidence before any candidate was advanced or rejected.
New York City (Local Law 144), Illinois (AIVEA), and the European Union (EU AI Act) are among the most specific. Canada's PIPEDA, California's CPRA, and Maryland's facial-recognition law add related requirements. The landscape is expanding quickly — employers should monitor their jurisdictions and consult local counsel.
In many jurisdictions, explicit consent is not required, but prior notice often is. Illinois's AIVEA, for example, requires employers to notify candidates before AI analyzes their interview. Consent requirements depend on local law and how the data is processed. Always disclose AI use clearly in your hiring process — it is both good practice and increasingly a legal baseline.
A bias audit is an independent statistical review testing whether an AI tool produces different outcomes for protected groups. NYC Local Law 144 requires annual audits for covered employers. Criba's scoring is tied to observable, job-relevant responses rather than demographic proxies, and its evidence-linked outputs can support an auditor's review. Whether a formal audit is required depends on your location and tool configuration.
Opt-out rights vary by jurisdiction. Illinois requires employers to offer a reasonable alternative assessment if a candidate objects to AI analysis. Even where not legally required, offering an alternative is a defensible practice. Build an alternative pathway into your process and document it — Criba does not restrict you from running parallel assessment tracks.
Yes. The EU AI Act classifies AI systems used for recruitment and candidate evaluation as high-risk. This triggers requirements for human oversight, transparency toward candidates, bias monitoring, and registration in an EU database. Employers deploying AI tools for EU-based hiring should work with legal counsel to map those obligations to their specific toolset and workflows.
No — nothing on this page or in Criba's product is legal advice. The regulatory environment for AI hiring tools is evolving, and obligations differ by jurisdiction, employer size, and use case. This page is provided for general informational purposes only. Always consult qualified employment or data-privacy counsel before deploying AI screening tools in your hiring process.