AI in HR Statistics 2026
By Axis Intelligence Research
Co-authors: Sarah Mitchell (AI & Machine Learning) and David Park (Careers & Workforce) | Last updated: August 9, 2026 | License: CC BY 4.0
Thirty-nine percent of organizations have deployed AI somewhere in their HR function, and recruiting is the single largest use case at 27%. But only 12% of HR professionals working in states that regulate employment AI have adjusted their policies to comply — a deployment-to-governance gap that Axis Intelligence Research measures at 66.1 on a 100-point scale.
Quick Answer
As of the December 2025 survey field date, 39% of organizations had adopted AI in their HR functions, with a further 7% planning launch during 2026, according to SHRM’s State of AI in HR 2026 (n=1,722 completed responses). Recruiting is the concentration point at 27% of organizations. Governance has not kept pace: Axis Intelligence Research’s HR AI Governance Gap Index (HRGGI™) reads 66.1 at its August 2026 baseline, driven by an 88-point compliance-inaction component. Independent of survey data, the New York State Comptroller found that New York City’s bias-audit regulator identified non-compliance at one-seventeenth the rate an independent review of the same 32 companies produced.
Key Findings
- 39% of organizations had AI deployed in HR as of December 2025, and 7% more planned to launch in 2026, per SHRM’s State of AI in HR 2026 survey of 1,722 HR professionals.
- Recruiting is HR’s largest AI use case at 27% of organizations, ahead of HR technology (21%), learning and development (17%), and employee experience (14%), according to the same SHRM survey.
- Axis Intelligence Research estimates the HR AI Governance Gap Index (HRGGI™) at 66.1 out of 100 at its August 2026 baseline, combining five sourced governance-shortfall components from SHRM’s 2026 data.
- The New York State Comptroller identified at least 17 instances of potential Local Law 144 non-compliance among 32 companies that New York City’s own regulator had cleared with a single finding, per audit report 2024-N-6 issued December 2, 2025.
- AI does not appear among the three most common business functions for AI deployment in nationally representative federal data, where Sales and Marketing (52%), Strategy and Business Development (45%), and IT (41%) lead among adopting firms, per U.S. Census Bureau working paper CES-26-25.
Source: Axis Intelligence Research — CC BY 4.0
How Many Organizations Use AI in HR in 2026?
The single most defensible number in this field comes from SHRM’s State of AI in HR 2026 report, fielded through the organization’s Voice of Work Research Panel between December 5 and December 23, 2025. Of 1,908 HR professionals who began the survey, 1,722 completed it. Data is unweighted.
Within that sample, 39% of organizations had AI adopted in their HR functions and 7% intended to launch during 2026 — the arithmetic behind SHRM’s headline framing that 46% of organizations expect to use AI in HR this year. A further 23% had AI deployed elsewhere in the organization but not in HR, putting total organizational AI use at 62%. Thirty-one percent had no AI plans at all.
| Adoption status | Share of organizations | As of | Source |
|---|---|---|---|
| AI adopted in HR function | 39% | Dec 2025 | SHRM, State of AI in HR 2026 |
| Plan to launch AI in HR during 2026 | 7% | Dec 2025 | SHRM, State of AI in HR 2026 |
| AI deployed elsewhere, not in HR | 23% | Dec 2025 | SHRM, State of AI in HR 2026 |
| No AI initiatives planned | 31% | Dec 2025 | SHRM, State of AI in HR 2026 |
The residual matters as much as the headline. Fifty-four percent of organizations have adopted no form of AI in HR and have no plans to do so this year — meaning the share using AI for recruiting, HR’s busiest application, is exactly half the share using no HR AI at all.
AI Adoption in HR by Organization Size
Adoption tracks headcount with unusual consistency. Extra-large organizations — SHRM defines these as 5,000 employees or more — report 60% HR AI implementation. Midsize organizations (100 to 499 employees) report 35%. Small organizations (2 to 99 employees) report 33%.
| Organization size | Employees | AI implemented in HR | Source |
|---|---|---|---|
| Extra-large | 5,000+ | 60% | SHRM, State of AI in HR 2026 |
| Large | 500–4,999 | Not separately reported | SHRM, State of AI in HR 2026 |
| Midsize | 100–499 | 35% | SHRM, State of AI in HR 2026 |
| Small | 2–99 | 33% | SHRM, State of AI in HR 2026 |
The spread between extra-large and small organizations is 27 percentage points — a ratio of 1.82 to 1. Recruiting is the most common practice area at every size band; the divergence appears after that. Extra-large organizations skew toward learning and development, talent analytics, and talent management. Small and midsize organizations skew toward performance management, C-suite and board relations, and — in the smallest band — organizational design and change management.
Which HR Practice Areas Use AI
SHRM assessed 138 use cases across 16 HR practice areas. The twenty most common clustered in just six of them.
| HR practice area | Share of organizations using AI | Source |
|---|---|---|
| Recruiting | 27% | SHRM, State of AI in HR 2026 |
| HR technology | 21% | SHRM, State of AI in HR 2026 |
| Learning and development | 17% | SHRM, State of AI in HR 2026 |
| Employee experience | 14% | SHRM, State of AI in HR 2026 |
| Inclusion and diversity | ≤2% | SHRM, State of AI in HR 2026 |
| C-suite and board relations | ≤2% | SHRM, State of AI in HR 2026 |
| Compliance | ≤2% | SHRM, State of AI in HR 2026 |
The distribution describes transactional work: resume parsing, interview scheduling, job-ad programming. More advanced applications exist — candidate-to-job matching, personalized learning recommendations, decision support — but they sit in the tail.
Sarah Mitchell: The shape of this distribution is the finding, not the 39%. AI in HR has landed almost entirely where the output is a queue rather than a judgment — parse this, schedule that, draft the posting. Compliance sits at two percent or below, which is the practice area where an error carries the largest downside and where a well-specified model would have the most defensible ground truth to learn from. Vendors sell decision support. Organizations are buying document handling. That gap between the roadmap and the install base is the number to watch in the 2027 field.
What Do Federal Data Say About AI in HR Compared With HR Surveys?
A second, entirely independent measurement exists, and it does not agree with the first — for reasons that turn out to be instructive rather than contradictory.
The U.S. Census Bureau’s Business Trends and Outlook Survey added an AI supplement fielded November 17, 2025 through February 8, 2026, drawn from a sample of roughly 1.2 million businesses. Census researchers analyzed it in working paper CES-26-25, published April 2026. During the supplement’s reference period, 18% of firms used AI in a business function — rising to 32% on an employment-weighted basis, and expected to reach 22% within six months.
The paper’s function-level breakdown is where HR’s position becomes visible. Among adopting firms, the most common deployments were Sales and Marketing (52%), Strategy and Business Development (45%), and IT (41%). Fifty-seven percent of adopting firms use AI in three or fewer functions at all. HR does not appear in the top three.
| Measure | Value | Population | As of | Source |
|---|---|---|---|---|
| Firms using AI in a business function | 18% | All U.S. employer firms | Nov 2025–Jan 2026 | U.S. Census Bureau, CES-26-25 |
| Same, employment-weighted | 32% | All U.S. employer firms | Nov 2025–Jan 2026 | U.S. Census Bureau, CES-26-25 |
| Expected within six months | 22% | All U.S. employer firms | Nov 2025–Jan 2026 | U.S. Census Bureau, CES-26-25 |
| Adopters using AI in ≤3 functions | 57% | AI-adopting firms | Nov 2025–Jan 2026 | U.S. Census Bureau, CES-26-25 |
| Firms reporting AI-related employment decrease | 2% | All U.S. employer firms | Nov 2025–Jan 2026 | U.S. Census Bureau, CES-26-25 |
Census also publishes a rolling biweekly series separate from the supplement. Its May 2026 analysis of AI use at U.S. businesses put the national rate at 19.8% as of May 3, 2026, with Information at 39.7% and Finance and Insurance at 33.9%.
Why Axis Intelligence Research Does Not Merge These Two Figures
It would be simple to average SHRM’s 39% against Census’s 18% and publish a blended adoption rate. That number would be meaningless, and we are stating the reason rather than producing it.
The two instruments measure different universes with different denominators. SHRM surveys HR professionals — a population that skews toward organizations large enough to employ dedicated HR staff and toward SHRM’s own membership, with unweighted results. Census surveys employer firms of all sizes, weighted to represent the full economy, and its 18% counts firms using AI in any business function, not HR specifically. A firm with four employees and no HR department is in the Census denominator and cannot be in SHRM’s.
The compatible comparison is not the headline rate. It is the employment-weighted Census figure — 32% — against the fact that SHRM’s own size cut puts extra-large-organization adoption at 60%. Both series say the same structural thing: AI adoption concentrates where headcount concentrates. Published side by side, they corroborate. Averaged, they would describe nothing.
How Big Is the AI Governance Gap in HR?
Deployment is measurable. Governance is measurable too, and the two are not moving together. Axis Intelligence Research built the HR AI Governance Gap Index (HRGGI™) to put a single number on the distance between the two.
HRGGI™ is a weighted composite of five governance-shortfall components, each expressed as the percentage of the relevant population that is not covered by a given control. All five inputs come from SHRM’s State of AI in HR 2026 survey and are reproduced in the accompanying CSV with full provenance. The index runs 0 to 100, where 0 means every measured governance control is universally in place and 100 means none of them is.
HRGGI™ Components and Weights
| Component | Underlying finding | Gap value | Weight | Contribution |
|---|---|---|---|---|
| G1 — Policy absence | 49% of AI-using organizations have workforce AI-use policies | 51 | 0.20 | 10.20 |
| G2 — Policy inadequacy | 25% of policy-holders call their policy clear and future-proof | 75 | 0.15 | 11.25 |
| G3 — Regulatory unawareness | 57% of HR professionals in regulated states are unaware of applicable law | 57 | 0.25 | 14.25 |
| G4 — Compliance inaction | 12% in regulated states have implemented compliant policies and practices | 88 | 0.25 | 22.00 |
| G5 — No outcome measurement | 56% do not formally measure the success of AI investments | 56 | 0.15 | 8.40 |
Formula: HRGGI = Σ(gapᵢ × weightᵢ), where weights sum to 1.00.
August 2026 baseline reading: 66.1.
Weights are editorial and disclosed rather than statistically optimized. G3 and G4 carry the heaviest weight (0.25 each) because unawareness of a binding legal obligation and failure to act on a known one are the two shortfalls that convert directly into legal exposure. G2 is weighted below G1 because an imperfect policy still establishes an accountable position; no policy establishes none. G5 is weighted lowest because unmeasured ROI is a management failure rather than a compliance one.
The reading is stable under alternative weighting. An equal-weight variant (0.20 across all five) returns 65.4 — a 0.7-point difference. Removing G4 entirely and renormalizing the remaining four returns 58.8, which establishes that the index is not a single-component artifact.
Scope. HRGGI™ measures governance controls that HR professionals self-report, not governance controls a regulator would verify. G3 and G4 apply specifically to respondents working in the 19 most populous states that SHRM identified as having enacted employment-related AI law or regulation as of February 2026; the other components apply to the full AI-using sample. The index captures policy existence and stated compliance action, not policy enforcement or audit outcomes. It does not measure vendor-side controls, which sit outside the surveyed population entirely.
Citation: Axis Intelligence Research, HR AI Governance Gap Index (HRGGI™), August 2026 baseline, axis-intelligence.com/ai-in-hr-statistics/. Published under CC BY 4.0.
A second figure sharpens the picture. Multiplying the two policy components — 49% of AI-using organizations have a policy, and 25% of those describe it as clear and future-proof — yields an effective governed share of 12.25%. Axis Intelligence Research estimates that roughly one AI-using organization in eight operates under an AI policy its own HR function considers fit for purpose.
David Park: Sixty-six is a number that should be read against who carries the consequence. The HR professional filling out that survey is generally not the person who bought the tool. SHRM’s own data has legal and compliance leading AI governance at 37%, and 52% of organizations not involving HR in overall AI strategy in any form. So the function most exposed to a discrimination claim is the function least present at the procurement table. That is not a training problem. That is an org-chart problem, and it will show up in a docket before it shows up in a survey.
Is Anyone Actually Enforcing AI Hiring Laws?
Here the data stops being about intentions and starts being about outcomes — and the most useful evidence is a government audit of a government regulator.
The New York City Local Law 144 Detection Gap
New York City’s Local Law 144 of 2021 prohibits employers and employment agencies from using an automated employment decision tool unless it has been subject to a bias audit within the prior year, the audit summary is publicly posted, and candidates receive required notice. The Department of Consumer and Worker Protection enforces it, with civil penalties running from $500 to $1,500 per day. Enforcement began July 5, 2023.
On December 2, 2025, the New York State Comptroller published audit 2024-N-6 on DCWP’s enforcement of Local Law 144, covering July 2023 through June 2025. Three findings define the state of enforcement.
DCWP received two AEDT complaints across the entire 24-month audit period — one complaint every twelve months, in a jurisdiction whose employers screen candidates at national scale. DCWP did not investigate whether its complaint intake process was functioning.
DCWP surveyed the websites and bias audits of 32 companies and identified a single instance of non-compliance. Comptroller staff reviewed the same 32 companies and identified at least 17 instances of potential non-compliance.
| Reviewing body | Companies reviewed | Non-compliance instances found | Rate | Source |
|---|---|---|---|---|
| NYC DCWP | 32 | 1 | 3.1% | NYS Comptroller audit 2024-N-6 |
| NYS Comptroller | 32 | ≥17 | ≥53.1% | NYS Comptroller audit 2024-N-6 |
Axis Intelligence Research calculates a detection gap of 50.0 percentage points — a 17-to-1 ratio — between the regulator’s finding rate and the independent reviewer’s finding rate on an identical sample. Inputs: 1 ÷ 32 = 3.125%; 17 ÷ 32 = 53.125%; difference 50.0 points; ratio 17 ÷ 1. The audit attributes the gap to DCWP not applying the formal procedures the NYC Office of Technology and Innovation developed under an interagency memorandum, and to DCWP’s review not addressing all Local Law 144 requirements. The audit also records that DCWP officials, who lack technical expertise to evaluate AEDT use, did not consult OTI when making AEDT determinations.
This is a small sample and should be read as one — 32 companies is not a census of New York City employers. But it is the only sample in existence where the same set of employers was reviewed twice by two independent bodies, and the divergence is not marginal.
The 2026 Regulatory Calendar Moved
Three things happened to employment-AI law between January and August 2026, and two of them ran in the opposite direction from where the field expected.
Illinois tightened. House Bill 3773, enacted as Public Act 103-0804, took effect January 1, 2026. The Illinois General Assembly bill record makes it a civil rights violation under the Illinois Human Rights Act for an employer to use AI that has the effect of subjecting employees to discrimination on the basis of protected classes, to use zip codes as a proxy for protected classes, or to fail to provide notice that AI is in use. Covered decisions span recruitment, hiring, promotion, renewal, training selection, discharge, discipline, and tenure. The Illinois Department of Human Rights was directed to adopt implementing rules on notice timing and means; those proposed rules were withdrawn, and as of late July 2026 no revised timeline had been published. The statutory obligations have been in force throughout.
Colorado retreated, then rewrote. Senate Bill 24-205, signed in 2024 as the first broad-scope U.S. state AI law, was delayed from February 1, 2026 to June 30, 2026 by SB 25B-004, stayed by a federal magistrate order on April 27, 2026, and then replaced outright by Senate Bill 26-189, signed May 14, 2026 and effective January 1, 2027. The original high-risk and algorithmic-discrimination framework never took effect. Its replacement is a narrower automated-decision-making disclosure regime.
The European Union deferred. Under Regulation (EU) 2024/1689, standalone high-risk AI systems listed in Annex III — a category that expressly includes recruitment, task allocation, worker monitoring, promotion, and termination — were scheduled to become subject to the full high-risk regime on August 2, 2026. The Digital Omnibus on AI, proposed November 19, 2025, endorsed by the European Parliament on June 16, 2026 and given final Council approval on June 29, 2026, moved that date to December 2, 2027. Annex I embedded systems moved to August 2, 2028. Article 50 transparency obligations largely held their original schedule.
| Jurisdiction | Instrument | Status as of August 9, 2026 | Operative date |
|---|---|---|---|
| New York City | Local Law 144 of 2021 | In force, enforcement audited as ineffective | July 5, 2023 |
| Illinois | HB 3773 / P.A. 103-0804 | In force; implementing rules pending | January 1, 2026 |
| Colorado | SB 24-205 → SB 26-189 | Original law replaced before taking effect | January 1, 2027 |
| European Union | Reg. (EU) 2024/1689, Annex III | High-risk obligations deferred | December 2, 2027 |
The week this article publishes is the week after the EU’s original Annex III deadline would have bound every recruitment AI system placed on the European market. It did not arrive. That is the context in which SHRM’s finding — 57% of HR professionals in regulated states unaware of the law that already applies to them — should be read: the compliance calendar has moved four times in eighteen months, and the function expected to track it is not the function that owns AI strategy.
Does AI in HR Eliminate HR Jobs?
The displacement question has a federal answer and a survey answer, and for once they point the same direction.
The Bureau of Labor Statistics projects employment of human resources specialists (SOC 13-1071) to grow from 944,300 in 2024 to 1,002,700 in 2034 — a 6% increase, or 58,400 additional positions, against 3% growth for all occupations. About 81,800 openings are projected annually across the decade. Median annual wage was $72,910 in May 2024. Human resources managers are projected to grow 5% over the same period, with about 17,900 annual openings and a May 2024 median of $140,030.
| Occupation | SOC | Employment 2024 | Projected 2034 | Change | Median wage (May 2024) | Source |
|---|---|---|---|---|---|---|
| HR specialists | 13-1071 | 944,300 | 1,002,700 | +6% / +58,400 | $72,910 | BLS Occupational Outlook Handbook |
| HR managers | 11-3121 | Not stated here | Not stated here | +5% | $140,030 | BLS Occupational Outlook Handbook |
| All occupations | — | — | — | +3% | $49,500 | BLS Occupational Outlook Handbook |
SHRM’s respondents describe the same non-event from inside. At organizations where AI has been deployed, HR professionals reported slight job displacement (7%), some new jobs or roles (24%), shifts in workers’ job responsibilities (39%), and frequent upskilling or reskilling opportunities (57%). Seventy-seven percent said AI use had no impact on their own job security; 73% said it had none on their career prospects.
SHRM frames this as AI being 5.7 times more likely to shift responsibilities and three times more likely to create new roles than to displace jobs. Recomputing directly from the published component shares returns 5.6 and 3.4 respectively — close enough to confirm the framing, and worth stating precisely rather than repeating a rounded multiple.
Census’s independent finding is blunter still: AI-related employment decreases occurred in 2% of firms, and 66% of AI users rely on it solely to augment tasks.
Applying SHRM’s 39% organizational adoption rate to the BLS specialist population produces an order-of-magnitude estimate: Axis Intelligence Research estimates that approximately 368,000 U.S. human resources specialists work in organizations that have deployed AI in the HR function (944,300 × 0.39 = 368,277). This is a cross-population estimate, not an observation. SHRM’s unweighted sample skews toward larger employers, so the true figure is plausibly lower.
Do Job Candidates Accept AI in Hiring?
The demand side of this market has an opinion, and it has been consistent across the only large-sample public survey available.
Pew Research Center surveyed 11,004 U.S. adults December 12–18, 2022, and published AI in Hiring and Evaluating Workers in April 2023. Americans opposed AI making a final hiring decision by 71% to 7%, with 22% unsure — roughly ten to one against. Views on AI reviewing applications were more divided: 41% opposed, 28% in favor, 30% unsure. On using AI-collected performance information to decide whether someone should be fired, opposition ran 55% to 14%. Two-thirds of adults said they would not want to apply for a job at an employer using AI to help make hiring decisions.
[older data] — Pew has not published an updated large-sample instrument on this specific question set. It is included because no comparably sized, methodologically transparent public-attitude dataset on AI in hiring has replaced it, and because the direction of the finding has not been contradicted by any subsequent primary survey we could locate.
| Application | Oppose | Favor | Not sure | Source |
|---|---|---|---|---|
| AI makes final hiring decision | 71% | 7% | 22% | Pew Research Center, April 2023 |
| AI reviews job applications | 41% | 28% | 30% | Pew Research Center, April 2023 |
| AI performance data informs firing | 55% | 14% | — | Pew Research Center, April 2023 |
SHRM’s 2026 data closes the loop from the employer side. Asked to imagine a world with no technical barriers to full HR automation, 72% of HR professionals said non-technical barriers would still prevent it — and within that group, 87% named the preferences of HR’s own customers, meaning employees, leadership, and job applicants. Legal or regulatory barriers came second at 57%. Cost-effectiveness came third at 35%.
Candidate resistance is not a soft constraint on AI in HR. In the survey data, it is the hardest one.
Methodology
Collection. All figures in this article were retrieved from named primary sources between August 7 and August 9, 2026. Every quantitative claim in the prose exists as a row in the accompanying CSV with source organization, source document, URL, retrieval date, and a primary-source flag.
Primary sources. SHRM, The State of AI in HR 2026 (survey fielded December 5–23, 2025; 1,908 started, 1,722 completed; unweighted; ANOVA where relevant). U.S. Census Bureau, Center for Economic Studies working paper CES-26-25, The Microstructure of AI Diffusion (April 2026; BTOS AI supplement fielded November 17, 2025–February 8, 2026; sample approximately 1.2 million businesses). U.S. Census Bureau, AI Use at U.S. Businesses (May 2026). U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, human resources specialists and human resources managers (2024–34 projections; May 2024 wage data). Office of the New York State Comptroller, audit 2024-N-6 (issued December 2, 2025; scope July 2023–June 2025). New York City Department of Consumer and Worker Protection, Local Law 144 program page. Illinois General Assembly, HB 3773 bill record (Public Act 103-0804). Pew Research Center, AI in Hiring and Evaluating Workers (n=11,004; fielded December 12–18, 2022).
Secondary attribution. The Colorado legislative sequence (SB 24-205 → SB 25B-004 delay → April 27, 2026 enforcement stay → SB 26-189) and the European Digital Omnibus vote records are drawn from legal-tracker reporting corroborated across multiple independent firms, and are flagged is_primary = no in the CSV. The underlying statutes and the EU regulation itself are primary instruments; the procedural dates are reported.
Synthesized figures. Three figures in this article are calculated by Axis Intelligence Research rather than reported by a source: the HRGGI™ reading (formula, components, and weights disclosed above), the Local Law 144 detection gap (1 ÷ 32 versus 17 ÷ 32), and the estimated HR specialist headcount in AI-adopting organizations (944,300 × 0.39). All three show their inputs inline. All arithmetic was verified computationally before drafting.
Restraint. SHRM’s 39% and Census’s 18% are published side by side rather than merged, because the two instruments use incompatible denominators and weighting. See the section above for the full reasoning.
Scope. Survey-based adoption figures reflect self-report by HR professionals and business respondents, not audited system inventories. SHRM’s data is unweighted and skews toward organizations with dedicated HR functions. The Local Law 144 comparison rests on a 32-company sample. Pew’s public-attitude data predates the current generation of generative AI hiring tools and is flagged accordingly.
About This Dataset
Title: AI in HR Statistics 2026 — Adoption, Governance, Regulation, and Workforce Data
Publisher: Axis Intelligence Research
Temporal coverage: December 2022 – August 2026
Spatial coverage: United States; European Union (regulatory)
Format: CSV, UTF-8, one observation per row, ISO 8601 dates
License: CC BY 4.0 — free to share and adapt with attribution
Provenance columns: metric, value, unit, as_of_date, geography, segment, source_org, source_document, source_url, retrieved_date, is_primary, axis_calculated, method_note
How to Cite
APA: Axis Intelligence Research. (2026, August 9). AI in HR statistics 2026: Adoption, governance gaps, and the enforcement data. Axis Intelligence. https://axis-intelligence.com/ai-in-hr-statistics/
MLA: Axis Intelligence Research. “AI in HR Statistics 2026: Adoption, Governance Gaps, and the Enforcement Data.” Axis Intelligence, 9 Aug. 2026, axis-intelligence.com/ai-in-hr-statistics/.
Chicago: Axis Intelligence Research. “AI in HR Statistics 2026: Adoption, Governance Gaps, and the Enforcement Data.” Axis Intelligence, August 9, 2026. https://axis-intelligence.com/ai-in-hr-statistics/.
To cite the index specifically: Axis Intelligence Research, HR AI Governance Gap Index (HRGGI™), August 2026 baseline, axis-intelligence.com/ai-in-hr-statistics/.
Last updated: August 9, 2026
Frequently Asked Questions
Does an applicant tracking system with resume parsing count as an AEDT under NYC Local Law 144?
Not automatically. The law reaches computational processes that issue a simplified output — a score, classification, or recommendation — used to substantially assist or replace discretionary decision-making. Tools that transcribe, translate, or capture data without scoring or recommending fall outside the definition. The practical test is whether the tool ranks or filters candidates in a way a human then relies on. A parser that extracts fields into a database is different from a parser that assigns a match score and sorts the queue.
If our headquarters is outside New York City, does Local Law 144 still apply to us?
It applies based on where the candidate resides, not where the employer sits. An employer evaluating a candidate who lives in any of the five boroughs is in scope, including for a fully remote role. This is why the law functions as a de facto national standard for any organization recruiting nationally.
Which HR practice areas are least likely to have AI deployed, and why does that matter?
Inclusion and diversity, C-suite and board relations, and compliance each sit at 2% or below in SHRM’s 2026 data. Compliance is the notable absence: it is the practice area where AI-assisted monitoring would most directly reduce the regulatory exposure that AI deployment elsewhere in HR creates. Organizations are automating the activity that generates legal risk faster than the activity that detects it.
Does HB 3773 require Illinois employers to conduct a bias audit?
No. Illinois took a different structural approach from New York City. HB 3773 imposes liability for discriminatory effect and requires notice when AI is used in covered employment decisions, but it does not mandate a periodic independent audit as a precondition of use. It also prohibits zip codes as a proxy for protected classes — a provision aimed squarely at geographic features that correlate with race in training data. Employers subject to both regimes need the audit for New York City candidates and the notice and non-discrimination controls for Illinois ones.
What happened to the August 2, 2026 EU AI Act deadline for recruitment tools?
It moved to December 2, 2027 for standalone Annex III high-risk systems, which includes recruitment, task allocation, worker monitoring, promotion, and termination tools. The Digital Omnibus on AI, endorsed by the European Parliament on June 16, 2026 and approved by the Council on June 29, 2026, made the change. Article 50 transparency obligations were largely unaffected. The deferral was attributed to incomplete harmonized standards and unfinished designation of national competent authorities, not to a decision to abandon the requirements.
Who owns AI governance inside the HR function?
Usually not HR. In SHRM’s 2026 data, legal and compliance functions lead AI governance and oversight at 37% of organizations. HR came closest to leading on upskilling and reskilling at 28%, still just behind cross-functional task forces at 29%. Fifty-two percent of organizations do not involve HR in overall AI strategy and vision in any capacity, direct or collaborative. HR professionals themselves largely endorse this: only 16% thought HR should lead change management for AI adoption, and 15% thought HR should lead employee training on AI tools.
Why do so many organizations deploy AI in HR without measuring whether it works?
Fifty-six percent of HR professionals in SHRM’s sample said they do not formally measure the success of AI investments at all, and only 16% use a return-on-investment metric of their own. Where measurement does happen, the top metrics are productivity, cost savings, decision quality, and employee satisfaction — three of which are difficult to attribute to a single tool inside a function whose outputs are shared across the business. The measurement gap and the governance gap have the same root: HR is generally deploying tools it did not select against objectives it did not set.
What is the single largest barrier to AI adoption among organizations that have not deployed it?
Not cost, and not risk. Sixty-seven percent of non-adopting organizations cited a lack of awareness of AI’s capabilities — uncertainty about where to start or how it would help. Technical concerns including transparency, accuracy, and insufficient internal data came second at 49%. Customer preferences, meaning the loss of human contact with employees and applicants, came third at 42%. Lack of resources came fourth at 38%. The largest barrier is the one most easily removed by education, which suggests adoption rates have further to climb before any structural ceiling is reached.
Is AI reducing the number of HR jobs?
Not in the data available. BLS projects HR specialist employment to rise 6% from 2024 to 2034 — faster than the 3% average for all occupations — reaching 1,002,700 positions. Census found AI-related employment decreases in 2% of firms overall, with 66% of AI users deploying it solely to augment tasks. SHRM’s respondents at AI-deploying organizations reported job displacement at 7% against shifted responsibilities at 39% and upskilling opportunities at 57%. The observable effect so far is task recomposition inside stable headcount, not contraction.
Related Axis Intelligence Research
- AI Statistics 2026 — macro adoption, market size, and the enterprise value gap
- AI Job Displacement Statistics — what the labor-market data shows and what it does not
- AI in Legal Statistics 2026 — the closest comparable professional-services vertical
- AI Spending Statistics 2026 — budget allocation and the enterprise return picture
- Tech Layoffs Statistics 2026 — workforce reduction data in the sector driving AI adoption
- Agentic AI Statistics — autonomous systems entering recruiting workflows