Artificial intelligence is no longer a novelty in the hiring process: It is close to standard practice. According to the World Economic Forum, roughly 90% of employers now use AI to filter or rank résumés. The legal rules on AI in hiring, however, are much less settled. Federal regulators have pulled back, states have stepped in and employers are left with a patchwork of rules that are easy to overlook until a problem arises.
How Employers are Using AI in Hiring
Most employers are not using a single “AI hiring tool” but are using a collection of technologies across the hiring process. Common examples include:
- Sourcing platforms that search for potential candidates who are not actively job-hunting and encourage them to apply;
- Résumé review software that scores applications and highlights top prospects from a high volume of submissions;
- Video platforms through which candidates answer interview questions on camera at a time convenient to them; and
- Software that evaluates interview responses and programs that automatically sort candidates by fit.
Employer use of AI is driven, at least in part, by the increased AI usage by applicants. Applicants can now generate a near-endless supply of résumés and cover letters, which can overwhelm recruiters and push businesses to rethink how they recruit and screen candidates. Because a résumé can be tailored by AI to fit almost any role, it has become a less reliable signal of whether a candidate can actually do the job. As a result, some reports indicate that many employers are shifting toward skills-based hiring, which evaluates applicants on abilities (such as interpersonal skills, problem-solving, expertise in a particular field) rather than on employment history or formal qualifications. The result is something of an arms race, with AI on both sides of the hiring table.
The Federal Retreat and the State Patchwork
During the Biden administration, the Equal Employment Opportunity Commission launched an initiative on AI and algorithmic fairness, issuing guidance on the disparate impact AI tools could have on protected groups. That approach began to change in April 2025, when Executive Order 14281 directed federal agencies to deprioritize enforcement of disparate impact liability. In September 2025, an internal EEOC memorandum directed staff to administratively close pending charges based solely on disparate impact and to issue right-to-sue letters, leaving charging parties to pursue those claims on their own in federal court.
Last December, Executive Order 14365 announced a federal goal of a “minimally burdensome” national AI policy. It also signaled that the federal government intends to challenge conflicting state AI regulations. The retreat has continued into 2026: in June, the Department of Justice’s Office of Legal Counsel concluded that the EEOC’s disparate impact guidelines are unconstitutional.
Some states have moved in the other direction. California and New Jersey adopted regulations prohibiting discriminatory automated decision systems, with New Jersey’s rules reaching disparate impact caused by automated employment decision tools even where employers rely on third-party vendors and lack discriminatory intent. Illinois amended its Human Rights Act, effective Jan. 1, 2026, to require notice of AI use in employment decisions.
Colorado recently passed a law, scheduled to take effect Jan. 1, 2027, that requires employers to explain an automated tool’s role in an adverse decision within 30 days and offer human review. The state attorney general is still developing implementation of the rules, and forthcoming regulations and guidance will likely shape what the law ultimately requires of employers.
While these are only some of the states that have begun to expand policies, we can expect to see more. This area of the law is evolving quickly, and employers should monitor implementing regulations and agency guidance, which will continue to flesh out what these laws require in practice.
What About The DMV?
Maryland, the District of Columbia and Virginia do not yet have laws that broadly govern AI hiring tools. Lawmakers in each jurisdiction have proposed such laws, but none has been enacted.
Maryland lawmakers in 2024 introduced House Bill 1255 and Senate Bill 957. Both bills would have limited employers' use of automated employment decision tools and required employers to notify applicants within 30 days of using one.
In the District of Columbia, then-Attorney General Karl Racine sent the D.C. Council the Stop Discrimination by Algorithms Act in 2021. It would have barred businesses from using algorithms that discriminate in employment. It also would have required covered entities to tell individuals how they use algorithms and to audit those algorithms for discrimination, and it would have created a private right of action. A version was reintroduced in 2023. To date, the bill has not become law.
In Virginia, the General Assembly passed in 2025 House Bill 2094, the High-Risk Artificial Intelligence Developer and Deployer Act. It would have required deployers of “high-risk” AI systems used for consequential decisions, including employment decisions, to manage risk, conduct impact assessments and make disclosures to consumers. Then-Gov. Glenn Youngkin vetoed the bill in March 2025, and the House sustained it.
In 2026, House Bill 999 would have made it an unlawful discriminatory practice under the Virginia Human Rights Act to use an automated decision system for employment decisions that results in discrimination or an unlawful disparate impact. It also would have required employers to tell affected applicants that a system was used and to assess the system for bias every year. That bill did not pass the House, and a separate 2026 bill on automated decision systems in employment decisions also failed.
While none of these proposals became law, taken together, they show that lawmakers across the region remain interested in regulating AI in hiring, so employers operating in all three jurisdictions should keep an eye on future sessions.
Key Legal Risks for Employers
Despite the Trump administration’s position and the absence of AI-specific hiring laws in many states, disparate impact liability remains part of Title VII itself, and it will stay the law unless and until Congress amends the statute or the courts declares it unlawful. Applicants can still file charges, obtain right-to-sue letters and bring disparate impact claims in federal court even if the EEOC itself declines to pursue them. In addition, state laws continue to apply even as federal enforcement steps back, exposing employers to liability in each state where they recruit or employ workers.
What Employers Can Do
Employers can start by inventorying every tool that screens, scores, ranks or otherwise influences hiring, including AI features embedded in broader HR platforms. Keeping good records is imperative, and failing to keep adverse-impact data may allow a plaintiff or agency to infer adverse impact from the gap. Such an inventory will also be useful if the employer becomes responsible for disclosing AI use under state law.
While some states do not mandate bias audits, proactive audits can provide valuable compliance evidence and litigation defenses. Vendor contracts should address bias testing and methodology, indemnification, and audit rights, and should secure the employer’s access to its own applicant flow and scoring data.
Employers also can incorporate meaningful human review of automated outputs, particularly rejections and borderline assessments. It is also important to update applicant notices to explain how to request reasonable accommodations if an AI tool could disadvantage individuals with disabilities.
Federal deregulation should not lull employers into complacency. Proceeding carefully and knowing what their technology is doing should save employers time and money in the long run and may help them avoid unnecessary litigation. Miles & Stockbridge's labor and employment lawyers are available to help employers assess their AI hiring tools and address compliance questions.
Opinions and conclusions in this post are solely those of the author unless otherwise indicated. The information contained in this blog is general in nature and is not offered and cannot be considered as legal advice for any particular situation. The author has provided the links referenced above for information purposes only and by doing so, does not adopt or incorporate the contents. Any federal tax advice provided in this communication is not intended or written by the author to be used, and cannot be used by the recipient, for the purpose of avoiding penalties which may be imposed on the recipient by the IRS. Please contact the author if you would like to receive written advice in a format which complies with IRS rules and may be relied upon to avoid penalties.