What AI Can (and Can't) Do for Talent Acquisition Teams

Every Talent Acquisition Leader has fielded the question by now. It comes from a senior leader who just sat through a conference session or read a think piece: "What are we doing with AI?" The expectation is usually ambitious.

The reality is more nuanced. AI has gotten genuinely useful in certain parts of the hiring workflow. It has also disappointed in ways that are predictable in hindsight. And there is now a growing layer of regulation that TA teams need to understand before they go much further.

What AI Is Actually Good At in Recruiting

There are real productivity gains available to TA teams from AI, but they are concentrated in specific stages of the workflow, not spread evenly across everything. The areas where AI is delivering the most consistent value tend to involve high volume, document generation, repetitive coordination tasks, or synthesis of large amounts of information. Think: finding candidates, generating tools to help evaluate candidates, managing scheduling logistics, keeping applicants informed, resources for your recruiters, and capturing what happened in an interview in a structured way.

Where AI tends to add less value is in anything that requires genuine judgment about people. Predicting whether a candidate will accept an offer, sensing a misalignment between what someone says they want and what the role actually demands, reading a room during a panel interview. Those things still belong to humans.

The important thing to understand is that AI's value in recruiting is not uniformly distributed. Some workflow stages are a strong fit. Others are not, or carry enough risk that deploying AI without the right safeguards creates more problems than it solves.

Where It Falls Short

The most visible limitation is bias. A Stanford-led study published in May 2026, covering four million job applications, found that AI screening tools used by Fortune 100 companies systematically disadvantaged Black and Asian applicants. SHRM research found that 19% of organizations using AI in hiring reported their tools overlooked qualified candidates.

Resume screening deserves particular attention here. It is one of the most common AI applications in recruiting, and one of the highest-risk from a discrimination and compliance standpoint. AI screening is most defensible when it supports a human reviewer rather than replacing one. Using it to generate a shortlist is a different proposition than using it to auto-reject.

Candidate perception is another real constraint. Survey data from ResumeBuilder found that roughly two-thirds of U.S. adults say they would avoid applying for jobs where AI makes hiring decisions. That matters more in competitive talent markets where employer brand and candidate experience drive whether people apply at all.

There is also the issue of resume signal degradation. As more candidates use AI tools to write and optimize their application materials, resumes become a less reliable differentiator. Some organizations are already shifting toward skills-based assessments and structured work samples as their primary selection signal, with AI handling volume earlier in the funnel.

The Compliance Layer You Cannot Ignore

Several jurisdictions now have laws that specifically regulate AI use in hiring, with disclosure requirements, consent obligations, bias audit mandates, and in some cases individual rights for candidates to request human review. The regulatory patchwork is only growing.

New York City's Local Law 144 is the most established. Employers using automated employment decision tools in New York City, including for remote roles, must conduct annual third-party bias audits and notify candidates that AI is being used before they are evaluated. Candidates can request an alternative process or opt out. A December 2025 audit by the New York State Comptroller found enforcement of the law has been "ineffective" to date, which signals that stricter scrutiny is likely coming, not that the law can be set aside.

Illinois has layered two laws on top of each other. The original AI Video Interview Act requires employers to notify candidates when AI analyzes video interviews, explain how the technology works, and obtain consent beforehand. Candidates can request that a human review their interview instead. Illinois HB 3773, which took effect January 1, 2026, extended those obligations to cover AI used in job postings, resume screening, and employment decisions more broadly. Employers must preserve notices and disclosures for four years, and liability extends to vendors and third-party recruiters acting on an employer's behalf.

California's amended FEHA regulations, in effect since October 1, 2025, cover any automated decision system that screens, scores, ranks, or recommends candidates, even where a human makes the final call. The regulations apply to employers with five or more California employees. The core obligation is non-discrimination: AI tools cannot produce disparate impact on protected classes, and employers are responsible for their vendors' tools, not just their own.

Maryland has a narrower law focused on facial recognition specifically, requiring candidate consent before any AI-powered facial recognition is used in an interview.

The practical implication across all of these is the same. Employers cannot outsource accountability to a software provider. If you are operating in or hiring into any of these jurisdictions, the right starting point is a clear-eyed audit of what AI tools are in your hiring workflow, which ones influence advancement decisions, and what you are actually disclosing to candidates.

Where to Start

For teams that have not made meaningful AI investments yet, the lowest-risk, highest-return places to begin are coordination and communication tasks. The compliance exposure is lower, the productivity gains are real, and the candidate experience impact is positive.

For teams already using AI in screening or candidate assessment, a compliance and bias review should come before any further expansion. Not as a reason to stop, but as a prerequisite for doing it defensibly.

The broader principle holds across all of it. AI works best in talent acquisition as decision support, not as a decision-maker. Every major AI hiring failure in recent years traces back to a system that was making or heavily influencing final decisions without sufficient human review. The teams getting consistent value from these tools are the ones that defined that boundary clearly before they deployed.

Knowing where to draw that line requires understanding your specific workflows, your candidate volume, and your legal exposure. That is a harder question to answer.

How Practical AI Advisor Can Help

At Practical AI Advisor, we work with HR and talent acquisition leaders to evaluate their current hiring workflows and identify where AI can realistically make a difference. That means looking at what you are actually doing today, where the friction points are, and what kinds of automation or AI tools fit your situation, your team, and your risk tolerance. The result is a clear, practical set of recommendations you can act on, not a vendor pitch or a generic AI roadmap.

If you are trying to figure out where to start, or whether the AI tools you are already using are the right ones, get in touch with us at mracine@practicalaiadvisor.com.