This article is in our Hiring Humans with Haley series, a series where we sit down with IQTalent technical recruiter Haley Crabiel to find out what's working, what's broken, and what forward-thinking teams are doing differently in engineering and technical hiring.
AI in hiring is everywhere. The policies governing it are not.
AI is reshaping every stage of the hiring process: how candidates apply, how recruiters screen, how decisions get made. Most organizations have already adopted the tools. Fewer have built the frameworks to use them responsibly. And almost none have kept pace with the regulations now arriving with real enforcement teeth.
The gap between adoption and governance is where risk lives. It's also where the most important conversations in talent acquisition are happening right now. This piece maps the landscape: what the rules require, where the confusion is, and what good actually looks like in practice.
AI adoption in hiring is outpacing governance—and the legal consequences are already catching up with companies that haven't built internal frameworks to match.The Policy Gap Is Real
Here's the paradox: 96% of recruiting teams now use AI in some capacity, according to the Criteria Corp 2025-2026 Hiring Benchmark Report. Yet the internal frameworks governing that use remain thin. A Resume.org survey from August 2025 found that while 78% of companies claim to have policies for ethical AI use in hiring, only 26% require human oversight for every candidate rejection. The policy exists on paper. The practice doesn't match it.
On the candidate side, things are equally murky. 40% of job seekers are already using AI to draft applications, per Insight Global's 2025 AI in Hiring Survey Report, yet they're doing it in a gray zone. Nobody told them it was okay. Nobody told them it wasn't.
That ambiguity is the problem.
Haley Crabiel, one of IQTalent's technical recruiters specializing in engineering hiring, put it simply when we sat down with her: "I say that AI is permitted. But what does that mean?"
It means something different to every person in the process.
Two Groups, Same Confusion
Recruiters are caught between wanting to embrace AI tools and not knowing how to advise the people they serve. They're being asked to be consultants on technology they haven't been trained on, for a process that hasn't been documented.
"Recruiters can be used as those consultants, but I also think recruiters could be educated on how to consult the client and the candidate themselves," Crabiel said.
That's a fair ask. It's also an unfunded one. As she noted: "That can only be fixed with some serious training, but that takes time and resources and money and all of the things. So it's just an overwhelming space for both the candidate and the company."
Candidates, meanwhile, are performing a kind of informal risk calculation every time they open ChatGPT before an interview. They want to put their best foot forward. They're not sure where the line is. "I think the candidate is nervous that it'll be looked at like cheating," Crabiel said. "So I think there also needs to be this mutual understanding of: this is okay."
That understanding doesn't exist yet. One in five U.S. professionals were secretly using AI in interviews as of early 2026, per data from Blind. "Secretly" being the operative word. They weren't disclosing it because they didn't know if they should. Meanwhile, 88% of hiring managers say they can detect AI-assisted applications, and 54% say they care, but there's no consistent response to it. Some penalize it. Some don't. Some haven't decided.
"One person is more for it than the other, and then it's like, what's the fine line there?" Crabiel said. "Where do we cross the cheating line? It's a very blurry space right now."
The Law Is Moving. HR Policy Isn't.
What makes this more than just an internal governance problem is that regulations are arriving with real teeth, whether organizations are ready or not.
NYC Local Law 144 has been in enforcement since mid-2023. Any employer using an automated employment decision tool (AEDT) to screen or rank candidates for jobs in New York City must conduct an annual independent bias audit, publish a summary of results publicly, and provide advance notice to candidates before using the tool. Fines apply. Complaints can be filed. The NYC Office of the State Comptroller audited enforcement as recently as December 2025. This is not theoretical.
Illinois now requires employers to notify job candidates when AI is used to analyze video interviews, obtain consent before AI evaluation occurs, and follow data retention rules. Provisions took effect in February 2026.
California expanded its civil rights framework in October 2025 to address AI-driven employment decisions, and prohibits hiring, firing, or promotion decisions using AI that result in discrimination against protected characteristics, as a potential violation of FEHA.
At the federal level, Title VII, the ADA, and the ADEA still apply, regardless of whether the decision was made by a human or an algorithm. As Bloomberg Law noted in May 2026: employers remain responsible for employment decisions, even when those decisions are informed by third-party technology. The use of a vendor doesn't shift legal accountability.
And the most sweeping framework hasn't fully landed yet. The EU AI Act classifies any AI system used in recruitment, candidate evaluation, or performance monitoring as high-risk. Full compliance obligations for employment-related AI take effect August 2, 2026. That means mandatory risk assessments, technical documentation, bias testing, human oversight requirements, and transparency disclosures for any organization whose AI tools touch candidates in the EU. Fines can reach €15 million or 3% of global annual turnover.
The extraterritorial reach matters: if your AI system screens a candidate for a role based in Berlin or Dublin, the EU AI Act applies to you, regardless of where your company is headquartered. For organizations with global recruiting operations, this isn't a regional concern. It's a baseline requirement.
Q: What laws govern AI use in hiring?
A: Multiple jurisdictions now regulate AI in hiring. NYC Local Law 144 requires annual bias audits and candidate disclosure for automated hiring tools. Illinois requires employer notification and consent for AI-analyzed video interviews. California prohibits AI-driven employment decisions that discriminate under FEHA. The EU AI Act, with full employment provisions taking effect August 2, 2026, classifies all recruitment AI as high-risk and requires risk assessments, bias testing, human oversight, and transparency. Federal employment laws (Title VII, ADA, ADEA) apply to AI-informed decisions. Employers using third-party AI tools remain legally responsible for outcomes.
The Patchwork Problem
Here's what makes the current environment particularly hard to navigate: there is no single standard. A company operating in New York, Illinois, California, and Germany is subject to four distinct regulatory frameworks, each with different requirements for disclosure, auditing, consent, and human oversight. As Bloomberg Law summarized, a single AI-driven tool may be subject to multiple, and sometimes inconsistent, legal standards.
Federal executive actions over the past year have signaled a lighter regulatory touch on AI, but they don't override state law, and private litigation has continued regardless. The Mobley v. Workday Inc. case, in which a federal court allowed disparate impact claims to proceed against an AI hiring platform, is a signal that courts aren't waiting for regulators to catch up.
"We all need to go to AI school or something," Crabiel said, only half joking. "Because it's just a whole thing."
She's right. And the absence of unified standards is the reason.
No unified federal standard means your AI hiring tools may be compliant in one state and a liability in another—here's what TA leaders need to know now.What Good Looks Like
Companies that are navigating this well aren't waiting for the rules to settle. They're building internal governance now. This is where Recruiting Operations Leaders have the most leverage, sitting at the intersection of the tools, the process, and the people affected by both.
A written AI use policy for the hiring process. Not vague principles. A document that specifies which tools are in use, what decisions they inform, and where human review is required. This should cover both the employer's use of AI and guidance for candidates on what's acceptable. IQTalent's CYBORG framework offers a starting point: mapping each stage of the recruiting process to determine where AI belongs and where human judgment needs to stay in the loop.
Bias audits before deployment, not after controversy. NYC Local Law 144 requires it. Other jurisdictions are moving in the same direction. Treat it as a baseline, not a compliance checkbox.
Candidate transparency as a default. Candidates have a right to know when AI is informing decisions about them. Disclosing this proactively, not burying it in a privacy policy, builds trust and reduces legal exposure. 79% of candidates want transparency when AI is used in hiring, per HireVue research. Give it to them.
Training for recruiters as advisors. The recruiter's role in an AI-enabled hiring process isn't just to use the tools. It's to explain them to hiring managers, to candidates, and to the organization's own HR and legal teams. That requires dedicated training, not assumptions. 72% of TA leaders plan to upskill teams on AI tools in the next 12 months, per Talent Board research. That investment should include compliance literacy, not just platform training.
Vendor contracts with accountability language. If you're using a third-party ATS, screening tool, or assessment platform, your contract should address audit rights, bias testing responsibilities, data use, and what happens if a tool produces discriminatory outcomes. Vendor claims of compliance are not the same as compliance.
The Recruiter's Role in All of This
The ambiguity in AI hiring policy isn't going to resolve itself quickly. Regulations are multiplying. Candidate behavior is evolving faster than company guidance. And internal HR teams are often the last to be resourced for this kind of infrastructure work.
That's exactly the context in which the recruiter, or a strategic recruiting partner, becomes more valuable, not less.
The companies that will navigate this best are the ones who have someone in the room who can translate: between what the law requires, what the tools can do, what candidates expect, and what the business needs. That's not an AI function. That's a human one.
The rules may be blurry. But the consequences of ignoring them are not.
IQTalent supports recruiting across global markets. Wherever your hiring happens, we're built to scale with it. Navigating AI compliance in your hiring process shouldn't require a law degree. Talk to IQTalent about building a recruiting strategy, through on-demand recruiting, RPO, and executive search with the right balance of AI efficiency and human oversight.
Next in the series: AI Didn't Replace My Expertise. It Made Me Better at Using It. — Haley Crabiel on how AI changed the way she shows up in hiring manager conversations.
Frequently Asked Questions
What laws regulate the use of AI in hiring?
Several jurisdictions now regulate AI in hiring. NYC Local Law 144 requires employers using automated employment decision tools to conduct annual bias audits, publish results publicly, and notify candidates before use. Illinois requires employer notification and consent before AI analyzes video interviews. California prohibits AI-driven employment decisions that result in discrimination under FEHA. The EU AI Act classifies all recruitment AI as high-risk, with full compliance obligations taking effect August 2, 2026. Federal laws—Title VII, the ADA, and the ADEA—apply to all employment decisions, including those informed by AI.
What is an automated employment decision tool (AEDT), and does it apply to my hiring process?
An automated employment decision tool is any software that uses machine learning, statistical modeling, or artificial intelligence to substantially assist or replace human judgment in employment decisions such as screening, ranking, or selecting candidates. Under NYC Local Law 144, employers using AEDTs for roles in New York City must comply with bias audit, disclosure, and notice requirements—regardless of where the employer is headquartered.
How should companies conduct bias audits for AI hiring tools?
Bias audits for AI hiring tools should be conducted by an independent third party before deployment, not after a complaint or incident. At minimum, audits should assess whether the tool produces disparate impact across race, sex, and other protected characteristics. NYC Local Law 144 requires these audits annually, with results published publicly. Leading organizations treat bias auditing as an ongoing operational requirement rather than a one-time compliance checkbox.
What should a written AI use policy for hiring include?
A written AI use policy for hiring should specify which tools are in use, what decisions each tool informs, and where human review is required. It should address both the employer's use of AI and guidance for candidates on what AI-assisted preparation is acceptable. Policies should include candidate disclosure practices, vendor accountability expectations, and a clear escalation process when the tool's outputs are questioned. Vague principles are not sufficient—the policy should be specific, operational, and reviewed at least annually as regulations evolve.
Are employers responsible for AI hiring decisions made by third-party vendors?
Yes. Employers remain legally responsible for employment decisions even when those decisions are informed by third-party AI tools. Using a vendor does not transfer legal accountability. Under Title VII, the ADA, and the ADEA, employers are liable for discriminatory outcomes regardless of the technology source. Vendor contracts should include audit rights, bias testing responsibilities, and explicit language about what happens if the tool produces discriminatory outcomes.