InterviewAgent.ai

AI screening · AI hiring compliance

AI hiring compliance for Local Law 144, EEOC rules and AI interview bias audits

The short answer

AI hiring compliance is the set of duties an employer takes on when software helps decide who advances: candidate notice and consent, disclosure that AI is involved, an independent bias audit where the law requires one, and records showing why each candidate was scored the way they were. Those duties sit with the employer, not the vendor. InterviewAgent.ai is built so the evidence exists by default: every candidate answers the same questions, every score links to a transcript, and a human makes every decision.

Last updated July 2026

Most AI hiring tools sell compliance as a checkbox, which is misleading in a way that costs employers money. Under NYC Local Law 144, the Illinois Human Rights Act and EEOC guidance, the duties land on the employer using the tool. No vendor can absorb that for you, and a vendor telling you otherwise is a reason for caution rather than comfort.

What a screening tool can do is make compliance cheap to prove. InterviewAgent.ai asks every applicant for the same role the same structured questions, scores each answer against a rubric your team defined, and keeps the transcript attached to the score. Candidates are told an AI conducts the interview and consent before it starts. The agent ranks and advances; it never rejects anyone on its own. When an auditor, a plaintiff's lawyer or your own general counsel asks why a candidate scored a 2, the answer is a sentence the candidate actually said.

Interview · score · rank · recruiter reviews · from $149/mo

Interview Studio

First-round interview

Candidate consented · AI-conducted

00:00 · AI Interviewer

Run the sample interview to watch the AI ask, follow up and score against your rubric.

Scored report

Rubric

The report assembles after the interview: overall score, rubric, highlights and a recommendation. You make the final call.

score

Highlights

Recommendation only · a recruiter makes the final decision

Ranked shortlist

Live, interactive · consent-first · no signup needed

Score /100 Ranked #1 of Transcript + highlights ready

Structured & consistent · bias-audited (EEOC / NYC Local Law 144) · you make the final call

ROLE-TAILORED RUBRIC-SCORED ATS-READY

Human-in-the-loop you decide

EEOC and LL144 bias-audited

Why it works

What your team gets with AI hiring compliance

Same questions, every candidate

Structured interviews for a role are identical in content and order, which removes the inconsistency that makes adverse impact hard to defend and easy to create.

Evidence behind every score

Each rubric score links to the transcript sentence that earned it, so you can show an auditor or a lawyer exactly what the decision was based on.

Consent and disclosure by default

Candidates are told the interview is conducted by AI and consent before it begins, which is the baseline Illinois requires and good practice everywhere else.

A human decides, always

The agent ranks and advances. It never auto-rejects, which keeps a person accountable for every employment decision you make.

What it handles

Interviewed, scored and shortlisted on autopilot

The agent invites each applicant, runs a role-tailored screening interview by voice or video, asks smart follow-ups, scores every answer against your rubric, and advances the strongest candidates into a ranked shortlist for your recruiters to review.

  • Runs identical structured questions for every applicant to a role
  • Discloses AI involvement and takes candidate consent before the interview
  • Scores against your rubric with the transcript attached as evidence
  • Supports bias auditing for EEOC and NYC Local Law 144
  • Keeps a recruiter making every advance or reject decision
SHORTLIST Interviewing
#1 Maya R. 92
#2 Devon L. 88
#3 Priya S. 81 Review
Rubric scored 48 screened this week

Why InterviewAgent.ai

One agent that runs the whole first round

Not a one-way video tool, not a six-figure assessment suite, and not a staffing agency. Interview, score, rank and hand off in one place, shaped to the roles and rubric you already hire on.

Interviews every applicant

A role-tailored screening interview runs by voice or video, with smart follow-ups, on the candidate's schedule. Applicants consent and are told they are speaking with AI, so no qualified person waits days for a first call.

Scores to your rubric

Every answer is scored against the same structured rubric, with transcripts and highlights, so candidates are compared consistently and the scoring stays bias-audited against EEOC guidance and NYC Local Law 144.

Ranks the shortlist

The strongest candidates are advanced into a ranked shortlist your recruiters review. The agent never auto-hires or rejects, it only surfaces who to talk to next, and your team makes every decision.

At a glance

What the main US rules require, and who carries the duty

Rule What it requires Who is liable
NYC Local Law 144 Independent bias audit within the prior 12 months, published summary, and 10 business days notice to candidates The employer or employment agency using the tool
EEOC guidance (Title VII) A selection procedure must not produce unjustified adverse impact, whoever built it The employer, including for a vendor tool
Illinois AI Video Interview Act Notice before the interview, an explanation of how the AI evaluates, and candidate consent The employer
Illinois HB 3773 (from Jan 1, 2026) Civil rights violation to use AI that discriminates, zip code named as an impermissible proxy The employer
Maryland HB 1202 Signed waiver before facial recognition creates a facial template The employer
Colorado SB 26-189 Broad AI duties, effective January 1, 2027, replacing the delayed SB 24-205 The deployer, which is the employer

Verified against the statutes in July 2026. Many roundups still print a June 30, 2026 Colorado date, which is out of date.

Does AI hiring software make my company compliant?

No, and this is the single most expensive misunderstanding in the category. Local Law 144 places the bias audit, the published summary and the candidate notice on the employer or employment agency that uses an automated employment decision tool. EEOC guidance is the same shape: if a selection procedure produces adverse impact, the employer answers for it even when a vendor wrote the algorithm. Buying a tool described as compliant does not move that duty.

What a tool genuinely changes is how much work compliance takes. If every candidate for a role got different questions from different interviewers, reconstructing why someone was rejected means interviewing your own recruiters months later. If the process was structured and recorded, the evidence already exists. Our guide to US state AI interview laws sets out each rule in detail.

What is a bias audit and who has to do one?

A bias audit under Local Law 144 is an independent evaluation of whether the tool produces different selection rates by sex and by race or ethnicity, including the intersections of the two. It has to be conducted within the previous 12 months by someone independent, which means not by you and not by the vendor, and a summary of the results has to be published where candidates can find it.

The duty attaches to the employer using an AEDT for a job in New York City, regardless of where the company is headquartered. If you are hiring for an NYC-based role from Austin, the rule reaches you. The definition of what counts as an AEDT is broader than most people expect, and a weighted spreadsheet can qualify: we cover the test in our explainer on automated employment decision tools.

  • Independent: not conducted by you or by the vendor selling the tool
  • Within the prior 12 months, and repeated annually
  • Tests sex, race or ethnicity, and the intersectional categories
  • A summary must be published publicly, not filed privately
  • Candidate notice at least 10 business days before the tool is used

What are the penalties for getting this wrong?

Local Law 144 penalties start at 500 dollars for a first violation, with up to 500 dollars more for each additional violation on the same day, and 500 to 1,500 dollars for each subsequent violation. The number that matters is not the unit price but the multiplier: each day the tool is used in violation is a separate violation, and failing to give notice is counted separately again. A quiet non-compliant quarter is not a 500 dollar problem.

Enforcement has been light so far, and that is changing. A New York State Comptroller audit released December 2, 2025 covering July 2023 to June 2025 found that the Department of Consumer and Worker Protection had reviewed 32 companies and identified one issue, while the Comptroller found at least 17 potential violations among those same companies. Around 75 percent of test calls to 311 about AEDTs were misrouted. The Comptroller made 13 recommendations and DCWP fully adopted 10 of them, so the reasonable planning assumption for 2026 is tighter enforcement, not continued quiet.

Can an AI reject a candidate automatically?

It can technically, and we deliberately do not allow it. The moment software rejects an applicant with no human in the loop, it is squarely an automated employment decision under Local Law 144 and a selection procedure under EEOC guidance, and the employer owns the resulting adverse impact. The design that lowers legal exposure also happens to hire better, because the candidate who interviews nervously but has done the exact job for six years stays visible.

Our AI interview agent ranks candidates and advances them to a recruiter with transcripts attached. A person decides who progresses and who does not, every time. Scores are evidence for that decision rather than a substitute for it.

What makes an AI screening round defensible?

Four things, and none of them are exotic. Ask every candidate for a role the same job-related questions. Score against a rubric written before you saw any candidates, with anchored levels rather than a vague impression. Keep the evidence, meaning the transcript behind each score. Keep a human making the decision, with the authority and the information to overrule the ranking.

Two things to avoid, both of which turn up in this category: do not let a tool infer personality or competence from facial expression or vocal tone, which is the weakest science in the field and the fastest route into Illinois and Maryland trouble, and do not score anything that is not job related. If you are writing the rubric now, our interview scorecard template gives you a structure that holds up.

Good questions

Questions about AI hiring compliance

Yes, in every US state, provided you meet the conditions that apply where you are hiring. New York City requires an independent bias audit and 10 business days of candidate notice, Illinois requires notice, explanation and consent for AI-analyzed video interviews, and Maryland requires a signed waiver before facial recognition. No state bans AI interviews outright.
It can. The rule follows the job, not the headquarters. If you use an automated employment decision tool for a role located in New York City, the bias audit, published summary and candidate notice duties apply to you even if your company sits in another state.
In Illinois, New York City and Maryland, yes, in the specific forms those laws describe. Everywhere else it is not strictly required, but disclosing it is the practical default: candidates find out anyway, and undisclosed AI screening is a reputational problem long before it is a legal one.
The employer, in almost every framework that matters. EEOC guidance treats a vendor tool as the employer's selection procedure, and Local Law 144 places the audit and notice duties on the employer or employment agency. Vendor indemnities may help commercially, but they do not move the legal duty.
We support it and cannot perform it for you, because Local Law 144 requires the auditor to be independent of both the employer and the vendor. We bias-audit the system supporting EEOC and Local Law 144 and provide the structured scoring data an independent auditor needs, and you commission the audit itself.
Costs vary widely with the number of tools and roles in scope, and no fixed public rate exists. The larger cost driver is usually data readiness: audits are far cheaper when scoring is already structured and per-candidate records exist, and expensive when an auditor has to reconstruct decisions from unstructured notes.

Explore more

More ways hiring teams screen with InterviewAgent.ai

Stop running first-round calls by hand. Put screening on autopilot.

Set your role and rubric and the agent interviews every applicant, scores each answer, and ranks a shortlist for your team. The agent advances candidates, your recruiters make every hiring decision.

See pricing

Role-tailored questions · bias-audited to EEOC and LL144 · human-in-the-loop