NYC AI Hiring Audits Just Got Real: What the Comptroller's Findings Mean for Employers in 2026
NYC AI Hiring Audits Just Got Real: What the Comptroller's Findings Mean for Employers in 2026
Three quarters of all complaints filed about AI hiring tools in New York City never reached the agency responsible for investigating them. And of the 32 companies a state auditor actually reviewed, the city found one potential violation. The state auditor found 17.
That is the gap at the center of the New York State Comptroller's December 2, 2025 audit of the Department of Consumer and Worker Protection's enforcement of Local Law 144. It is also why your compliance posture on AI hiring tools needs to change right now.
What LL144 Actually Requires
NYC Local Law 144 (effective July 5, 2023) applies to any employer or employment agency that uses an Automated Employment Decision Tool (AEDT) to screen candidates or rank employees for promotion in New York City. If that describes your organization, three obligations apply:
- Bias audit: An independent bias audit must be conducted annually, testing the tool for disparate impact across sex and race/ethnicity categories.
- Public disclosure: Employers must publish a summary of the audit results and the tool's intended use on their public-facing website.
- Candidate notice: Job candidates and employees must be notified — at least ten business days in advance — that an AEDT will be used, and must be given the opportunity to request an alternative selection process.
All three obligations, for every AEDT in scope, every time. This is not a best-practices framework. It is a legal requirement with daily penalties attached.
What the Audit Found
The Comptroller's audit, published December 2, 2025, reviewed DCWP's enforcement record from July 2023 through mid-2025. The findings were stark.
The most consequential problem: DCWP's primary complaint intake channel was broken. 75% of LL144 complaints submitted through the city's 311 system were misrouted — they went to other agencies or sat unresolved, never reaching DCWP at all. Enforcement was functionally blind to a large share of incoming complaints.
The disparity in findings tells the rest of the story. DCWP reviewed the same 32 companies the Comptroller examined and found 1 potential violation. The Comptroller's office found 17 in the same set. The gap wasn't investigative bad luck — it reflected structural weaknesses in DCWP's methodology: insufficient proactive monitoring and limited technical capacity to evaluate AI tools in any depth.
For employers, the implication is direct. The loose enforcement environment of the past two years is over. The audit has created institutional and political pressure on DCWP to demonstrate it can actually do the job.
What Changed
DCWP's public response to the Comptroller's findings includes specific operational commitments:
- 311 intake fix: DCWP will work with the city's 311 system to ensure LL144 complaints are correctly routed to the agency going forward.
- Cross-trained staff: Agency personnel will receive training to identify and evaluate AEDTs during employer reviews.
- Proactive monitoring: DCWP will move beyond complaint-driven enforcement and actively check employer websites for published audit summaries and required candidate notices.
- Direct employer interviews and tool demonstrations: Rather than reviewing documents alone, DCWP will conduct employer interviews and request live tool demonstrations to understand how AI tools are actually being used in hiring decisions.
- OTI partnership: DCWP will coordinate with the NYC Office of Technology and Innovation to strengthen its technical assessment capabilities.
Proactive monitoring is the most significant change. You no longer need a candidate complaint to trigger an inquiry. DCWP can — and has now committed to — find non-compliant employers on its own.
The Penalty Math
Do not treat LL144 violations as a manageable cost. The penalty structure is $500 to $1,500 per violation per day, and the compounding arithmetic is uncomfortable.
Consider a company using two AI screening tools — one for resume screening, one for structured interviews — with gaps across all three obligation categories (no current bias audit, no public disclosure, no candidate notice):
2 tools × 3 violation types × $1,500/day = $9,000/day
Over 30 days of non-compliance: $270,000.
That calculation assumes the maximum daily rate and simultaneous citation of all violation types. It does not include legal fees, remediation costs, or reputational exposure. But it illustrates what a focused enforcement action — one now backed by proactive monitoring and direct tool demonstrations — can generate.
2026 Action Checklist
If your organization uses AI tools to screen or rank candidates in New York City, these are your immediate priorities:
- Inventory every AEDT in scope. Resume screening tools, structured video interview platforms, automated scoring for promotion decisions — all potentially qualify. Know what you're running.
- Confirm your bias audit is current. Audits must be independent and conducted within the last twelve months. If yours is lapsed or has never been done, commission one now.
- Verify your public disclosure. The bias audit summary — including selection rate data by sex and race/ethnicity — must be live on your website. Check that it exists, is current, and is findable.
- Audit your candidate notice process. Written notice must go to candidates at least ten business days before an AEDT is used. Verify it is being sent, saved, and tracked consistently.
- Review tool architecture with legal counsel. Tools that use human-in-the-loop review — where AI informs but does not replace the final hiring decision — and that operate on transcript content only (not biometric voice or facial characteristics) carry meaningfully different AEDT exposure under LL144. Know where your tools sit on that spectrum.
- Prepare for a live demonstration request. DCWP has committed to requesting tool demos from employers. Your team should be able to walk an auditor through how your AI screening tool works, what outputs it produces, and how those outputs are used in hiring decisions.
- Document the full compliance chain. Audit dates, vendor certifications, candidate notice delivery records, disclosure publication history — in a proactive enforcement environment, your records are your defense.
- Put compliance on a recurring calendar. Annual bias audits, annual website disclosure reviews, and quarterly notice process checks should be standing agenda items, not one-time fixes after a scare.
The Bigger Picture
NYC LL144 is no longer an isolated requirement. Two additional state laws are active or approaching:
Illinois HB 3773 became effective January 1, 2026. It requires employers using AI in employment decisions to notify candidates, disclose the factors the AI evaluates, and provide a meaningful appeals mechanism. Unlike LL144, it applies to any Illinois employer — not just those in a specific city.
Colorado SB 24-205 takes effect June 2026. It imposes developer and deployer obligations around algorithmic discrimination and explainability for "high-risk" AI systems, a category that includes hiring tools.
The compliance landscape for AI in hiring is moving from patchwork to patterned. What NYC began in 2023, Illinois and Colorado are extending. If your organization operates across multiple states, managing these requirements jurisdiction-by-jurisdiction is no longer viable. A unified AI hiring compliance framework — covering notice, audit, disclosure, and human oversight requirements — is the only practical approach.
Act Before DCWP Knocks
The Comptroller's audit was a flare. DCWP's response commitments — proactive monitoring, cross-trained staff, direct employer interviews — are the signal that enforcement is real now, not theoretical.
The companies the Comptroller flagged in December 2025 didn't necessarily know they were non-compliant. Proactive enforcement closes that gap. It means the absence of a complaint is no longer a signal that you're fine.
Fix your LL144 posture now. Not when a DCWP investigator requests a tool demonstration, and not when a complaint finally routes correctly for the first time.
Sources
- NY State Comptroller audit, December 2, 2025 — https://www.osc.ny.gov/state-agencies/audits/2025/12/02/enforcement-local-law-144-automated-employment-decision-tools
- DLA Piper analysis, January 2026 — https://www.dlapiper.com/en-us/insights/publications/2026/01/critical-audit-of-nyc-ai-hiring-law-signals-increased-risk-for-employers
- NYC DCWP AEDT page — https://www.nyc.gov/site/dca/about/automated-employment-decision-tools.page
- DCI Consulting audit summary — https://blog.dciconsult.com/recommended-changes-to-local-law-144