A few days ago I posted on LinkedIn about a problem nobody in the business community wants to talk about: companies that didn’t build artificial intelligence but are quietly running their operations on it — real estate agents, attorneys, medical practices, device companies, financial advisors — using AI to write contracts, draft client communications, evaluate decisions, and generate work product, and then handing that work to clients as if a human did all of it, with no disclosure, no review, and no second set of eyes.

That post was about general AI liability. This one is more specific, and frankly more urgent. Because of every place AI has quietly moved into the modern business, the human resources department is the single most exposed, and almost nobody is talking about what that means for Employment Practices Liability Insurance (EPLI), general liability, or professional liability coverage.

Here’s the short version: your HR team has almost certainly already let an algorithm do the hiring, you probably aren’t telling candidates, you may be sitting on a discrimination claim you don’t even know about, and there is a very real chance your insurance won’t respond when the demand letter shows up.

Let’s break that down.

AI Is Already Running Your Hiring, Whether You Know It or Not

You don’t have a choice anymore about whether AI is involved in your hiring. You only have a choice about whether you know it’s happening and whether you’ve managed the risk.

The adoption numbers are staggering. In 2024 alone, AI-powered hiring tools processed more than 30 million applications in the U.S.1 By 2025, 43% of organizations used AI for HR tasks, up from 26% a year earlier.2 In tech, that adoption rate is 89%. In financial services, 76%. In healthcare, 62%.3 Roughly 99% of hiring managers in recent surveys say they use AI in some capacity in the hiring process.4 Workday alone disclosed in court filings that 1.1 billion applications were rejected through its software during the relevant class period in just one ongoing case.5

The tools include resume scanners that prioritize keywords, “virtual assistant” chatbots that screen candidates against predefined criteria, video interview platforms that analyze facial expressions and tone, predictive scoring engines, ranking algorithms, and increasingly autonomous “AI recruiter” agents that source, screen, and schedule with no human in the loop until the late stages.

If you run a business with more than a handful of employees and you post jobs online or use a major Applicant Tracking System (ATS), an AI is almost certainly making or “materially influencing” decisions about who gets seen and who gets rejected. Most owners and executives have no idea what it’s evaluating, how it was trained, or what its disparate impact profile looks like on protected classes.

That ignorance is the liability.

The Disclosure Problem: 70% of Candidates Are Never Told

Here is what makes this exposure so dangerous. The use of AI in hiring is widespread, but the disclosure of that use is almost nonexistent.

A recent Greenhouse survey of nearly 1,200 U.S. job seekers found that about 70% were never told the hiring process would involve being evaluated by AI. Roughly a fifth only found out when they started the interview.6 A separate study reported that less than 10% of candidates were clearly informed an AI was evaluating them, and combined with the “don’t know” group, 84.7% of applicants are walking into hiring processes blind to what’s reading their resume or scoring their facial expressions.7

That is a massive problem. Once a candidate is rejected — and let’s be honest, almost all of them get rejected — they have no idea whether a human read their materials, whether they were screened out by an algorithm, whether the algorithm was tested for bias, or whether they have a basis to challenge anything.

And under federal anti-discrimination law, that uncertainty is now the employer’s problem, not the candidate’s.

The EEOC Has Already Drawn the Line

The Equal Employment Opportunity Commission launched its agency-wide Artificial Intelligence and Algorithmic Fairness Initiative back in 2021.8 Since then, the agency has issued explicit guidance under both the Americans with Disabilities Act (May 2022)9 and Title VII of the Civil Rights Act (May 2023).10

The guidance is not ambiguous. The EEOC’s position is that an employer is responsible for discriminatory outcomes from an automated hiring tool even if a third-party vendor designed it, administers it, or scores it. The employer is on the hook if the tool produces a disparate impact on race, color, religion, sex, national origin, age, or disability. “We outsourced it” is not a defense.

The agency also signaled that employers may be liable for an agent’s actions, including software vendors who have been given authority to act on the employer’s behalf. That includes a vendor whose software is making the rejection decision.

In 2023, the EEOC settled its first AI hiring discrimination case, EEOC v. iTutorGroup, for $365,000. iTutorGroup had configured its recruiting software to automatically reject female applicants 55 and over and male applicants 60 and over. The smoking gun was an applicant who submitted two identical applications — one with her real birth date, one with a younger one — and got an interview only with the younger date.11 More than 200 applicants were affected. The consent decree runs for five years.12

That was the first. It will not be the last.

Mobley v. Workday: The Case That Should Terrify Every HR Department

If iTutorGroup was the warning shot, Mobley v. Workday is the cannon.

Derek Mobley, an applicant over 40, sued Workday alleging its AI applicant recommendation system rejected him from approximately 100 jobs based on age, race, and disability. He noticed something suspicious: rejection emails arrived outside business hours, sometimes within an hour of submission. No human was looking at his applications.

The case is being litigated in the Northern District of California. The headlines so far:

Allegedly widespread discrimination is not a basis for denying notice. — Judge Rita Lin

Here’s the part that matters for every business in America: in court filings, Workday represented that 1.1 billion applications were rejected through its system during the relevant period. The judge acknowledged that the collective “could potentially include hundreds of millions” of members.5

Workday is used by more than 11,000 organizations worldwide.18 Every one of those organizations is a potential defendant in spinoff litigation, because under the EEOC’s guidance and the court’s agency-theory reasoning, the employer using the tool is liable for the discriminatory outcome, not just the vendor.

And Workday is not alone. In March 2025, the ACLU filed a complaint against HireVue and Intuit on behalf of a deaf, Indigenous applicant whose AI video interview rejected her with feedback that she needed to “practice active listening.”19 A Stanford study published in 2026 tracked 3.4 million applicants across 1,700 job postings at 150 employers and found that AI screening tools were systematically disadvantaging Black and Asian applicants,20 and that 10% of applicants who applied to four different employers using the same vendor were rejected by all of them.21 The same algorithm rejecting the same person across unrelated employers. Imagine the class action on that one.

State and Local Laws Are Multiplying

While the federal government remains largely on the sidelines, states and cities are filling the gap, and the patchwork is brutal for any employer operating across state lines.

For an employer hiring across multiple states, the compliance load is significant, and the answer “we didn’t know we were even using AI” is not going to fly with a regulator or a plaintiff’s lawyer.

And Then There’s the Ghost Jobs Problem

Now layer on top of all of this another set of exposures that the same HR departments have created: the ghost jobs problem.

A “ghost job” is a publicly posted job listing that the employer has no current intent to fill. The data is, frankly, damning:

Government roles run the highest ghost-job rate at roughly 60%. Tech is around 48%.32

Now here’s the part most employers don’t see coming. The Columbia Law Review’s November 2025 analysis argues — convincingly — that ghost jobs already violate Section 5 of the FTC Act, which prohibits unfair or deceptive practices in commerce.32 FTC Chair Andrew Ferguson directed the agency in February 2025 to form a Joint Labor Task Force to address deceptive job advertising.33 Bills in California and Kentucky in 2025 would require employers to disclose whether a posting is for a real vacancy. Ontario, Canada has already enacted legislation taking effect in 2026.34

The legal theories already on the table against employers posting ghost jobs include fraudulent misrepresentation, negligent misrepresentation, false advertising under the Lanham Act, and unfair business practices under state UDAP statutes. And here is the real kicker: if a ghost job is posted and an AI is screening the responses, you have layered a disparate-impact discrimination exposure on top of a deceptive trade practices exposure on top of a potential wage-and-hour exposure if the posting was used to justify keeping current employees underpaid. One job posting; three categories of claim.

The Insurance Coverage Gap Nobody Is Talking About

This is where, as your insurance agent, I have to be blunt.

The first problem is that most small and mid-sized businesses don’t have Employment Practices Liability Insurance at all. They have General Liability, maybe Professional Liability if they’re in a professional services business, maybe a BOP. They view EPLI as optional. It is not optional. Once you have your second employee, you have EPL exposure, and once you start using AI to screen anyone — applicant, employee, or contractor — your exposure has multiplied by a factor most policyholders do not appreciate.

The second problem is that even if you carry EPLI, it may not respond the way you expect when an AI-driven discrimination claim comes in. Most EPLI policies were written before generative AI exploded into HR workflows. Carriers are increasingly adding exclusions or restrictive endorsements specifically targeting AI-driven decisions and algorithmic discrimination. In early 2026, major carriers including Chubb, Travelers, and Berkshire Hathaway received state regulatory approval to add explicit AI exclusions to general liability, D&O, and E&O policies. According to industry reporting, state regulators approved more than 80% of those requests, and in several states the exclusions began taking effect as early as January 2026.35

So when the demand letter arrives alleging that your AI-driven screening tool disparately impacted applicants over 40, here is what you need to ask:

  1. Does my EPLI policy define “wrongful act” broadly enough to capture algorithmic discrimination, or has it been silently endorsed to exclude AI-driven decisions?
  2. Does my GL or professional liability policy have a new AI exclusion that was added at the last renewal?
  3. If a third-party vendor was the proximate cause, do I have a contractual indemnity that’s actually enforceable, and is it backed by collectible insurance on the vendor’s side?
  4. What about the defense obligation? Even if coverage for indemnity is murky, will the carrier pay to defend?
  5. What’s the retention, what’s the limit, and is it shared with other coverages on a stacked tower?

Most insureds cannot answer any of these questions about their own policies. Most brokers haven’t asked the questions either. That is the gap.

A clean EPLI policy with the right endorsements, paired with a Cyber/Tech E&O or standalone AI liability policy where appropriate, is now a baseline for any business using AI in operations, and especially any business using it in hiring or HR. Specialized AI liability products are now being offered by carriers like Munich Re, with limits ranging from $2 million to $50 million.36 The market is moving fast. So is the litigation.

A Word to the Plaintiffs’ Bar

I’ll say what insurance brokers usually don’t say out loud: the EPL plaintiffs’ bar is sitting on top of one of the largest litigation opportunities of the decade and, with notable exceptions, has been slow to mobilize.

Consider the math. Workday’s own filings indicate 1.1 billion application rejections through its tools. The court has authorized notice to a collective that “could potentially include hundreds of millions.” That is one vendor. There are dozens of similar AEDTs in widespread use. There are 11,000+ Workday customer organizations alone, each of which may have its own derivative exposure under the agency theory the court has already accepted.

Layer in 93% of HR professionals admitting they post ghost jobs, and the Columbia Law Review forum arguing that ghost jobs themselves are already actionable under FTC Section 5 and existing UDAP statutes. Layer in the ACLU’s complaint against HireVue and Intuit for ADA and Title VII violations in AI video interviewing. Layer in the EEOC’s stated enforcement priority on algorithmic fairness. Layer in California’s October 2025 ADS regulations that explicitly make bias testing — or the lack of bias testing — relevant evidence in a discrimination claim.

Employment attorneys who specialize in plaintiff-side discrimination work should be opening offices, not closing them. The case theory is established. The defendants are well-funded. The classes are enormous. The insurance is increasingly being structured to not respond, which means the recovery comes from the defendant’s balance sheet, and the named plaintiffs and putative class members are out there in the millions, most of them never having been told an algorithm rejected them and many of them within ADEA, ADA, or Title VII protected classes.

This is not a hypothetical exposure. The court orders are already on the docket.

What Hendrickson Insurance Recommends Right Now

If you run a business — any size, any industry — and you use any third-party platform to post jobs or screen applicants, do the following this quarter, not next year:

  1. Audit your hiring stack. Identify every tool, ATS, scoring engine, chatbot, and video interview platform in the recruiting workflow. If you don’t know what the tool does or how it was trained, that is itself a finding.
  2. Disclose AI use to candidates. Whether your state requires it yet or not. The plaintiffs’ bar is going to use the absence of disclosure as evidence of intent.
  3. Get bias audits done. Even if you are not in NYC, follow the NYC LL 144 model. It is rapidly becoming the de facto national standard, and a documented audit is your best defense.
  4. Stop posting ghost jobs. Period. The data shows your HR team is almost certainly doing it. The exposure is no longer theoretical.
  5. Pull your EPLI policy off the shelf and read it with your broker, page by page. Look specifically for AI-related endorsements or exclusions, definitions of “wrongful act,” third-party vendor liability provisions, and the defense-cost language. If you don’t have EPLI, get it. Today.
  6. Review your GL, Professional Liability, D&O, and Cyber policies for new AI exclusions. Renewals in 2026 are quietly carrying these. Don’t sign without asking.
  7. Push the indemnity upstream. If you are using a third-party AEDT vendor, your contract should have a robust indemnification, backed by their insurance, and your broker should verify that their coverage actually responds to algorithmic discrimination.

The companies that get this right will sleep at night. The ones that don’t will become case law.

Where Hendrickson Insurance Fits In

Hendrickson Insurance is an independent Gulf Coast Florida agency — based in Sarasota and serving business owners across Bradenton, Tampa, St. Petersburg, Clearwater, Venice, and Lakewood Ranch. We specialize in complex commercial risk for businesses operating in AI-influenced industries, including technology, fintech, real estate, professional services, and healthcare. We engineer EPLI, Cyber, Technology E&O, and AI Liability programs that are built for the way modern businesses actually operate.

If you’ve read this article and don’t know whether your current coverage responds to an AI-driven employment claim, that is the conversation we should be having. Book a free coverage review — it costs nothing, and you’ll know exactly where you stand before your next renewal.

Do you know your agent?

References

  1. Reeya Khuarana, “AI in Hiring: Emerging Legal Developments and Compliance Guidance for 2026,” HR Defense (Proskauer Rose LLP), November 20, 2025. hrdefenseblog.com
  2. SHRM, “The Role of AI in HR Continues to Expand,” 2025 Talent Trends Report, 2025. shrm.org
  3. “121 AI in Recruitment and Hiring Statistics for 2026,” Novoresume, March 2, 2026. novoresume.com
  4. Insight Global, 2025 AI in Hiring Survey Report (Atomik Research, n=1,005 hiring managers). insightglobal.com
  5. Proskauer Rose LLP, “AI Bias Lawsuit Against Workday Reaches Next Stage as Court Grants Conditional Certification of ADEA Claim,” June 11, 2025. proskauer.com
  6. Greenhouse, 2026 Candidate AI Interview Report, PR Newswire, May 1, 2026. prnewswire.com
  7. Enhancv, “AI Hiring in 2026: Half of Job Seekers Were Rejected Without a Word,” April 2026 (n=1,066). enhancv.com
  8. U.S. EEOC, “EEOC Launches Initiative on Artificial Intelligence and Algorithmic Fairness,” October 28, 2021. eeoc.gov
  9. U.S. EEOC, ADA AI guidance, May 12, 2022. eeoc.gov
  10. U.S. EEOC, Title VII AI guidance, May 18, 2023. eeoc.gov
  11. Sullivan & Cromwell LLP, “EEOC Settles First AI-Discrimination Lawsuit,” August 16, 2023. sullcrom.com
  12. U.S. EEOC, “iTutorGroup to Pay $365,000 to Settle EEOC Discriminatory Hiring Suit,” September 11, 2023. eeoc.gov
  13. Holland & Knight LLP, “Federal Court Allows Collective Action Lawsuit Over Alleged AI Hiring Bias,” May 27, 2025. hklaw.com
  14. Norton Rose Fulbright, “Workday AI Lawsuit Receives the Greenlight to Proceed as a Collective Action,” June 4, 2025. insidetechlaw.com
  15. Civil Rights Litigation Clearinghouse, “Mobley v. Workday, Inc. 3:23-cv-00770 (N.D. Cal.).” clearinghouse.net
  16. allaboutlawyer.com, “Workday Class Action Lawsuit,” January 4, 2026. allaboutlawyer.com
  17. TheSlowAI, “Workday AI Hiring Lawsuit,” April 2026. theslowai.substack.com
  18. Workday, Inc., PR Newswire, September 11, 2025. prnewswire.com
  19. ClassAction.org, “AI Job Screening, Interview & Hiring Lawsuits,” February 4, 2026. classaction.org
  20. Stanford Institute for Human-Centered AI, “AI Hiring Tools Can Yield Racial Bias and Systemic Rejection,” May 2026. hai.stanford.edu
  21. Allwork.space, “Study Finds AI Hiring Tools May Be Rejecting The Same Candidates Across Multiple Companies,” May 2026. allwork.space
  22. NY State OSC, “Enforcement of Local Law 144, Automated Employment Decision Tools,” December 2, 2025. osc.ny.gov
  23. Mayer Brown LLP, “California Adopts New Employment AI Regulations Effective October 1, 2025,” September 2, 2025. mayerbrown.com
  24. K&L Gates LLP, “Illinois Anti-Discrimination Law to Address AI Goes Into Effect on 1 January 2026,” National Law Review. natlawreview.com
  25. Ogletree Deakins, “Colorado’s New AI Act Targets Automated Decision-Making for Consequential Decisions,” May 2026. ogletree.com
  26. Norton Rose Fulbright, “The Texas Responsible AI Governance Act.” nortonrosefulbright.com
  27. The Interview Guys, “Ghost Jobs Exposed,” November 12, 2025 (citing LiveCareer March 2025). theinterviewguys.com
  28. DAVRON, “Ghost Jobs & Misleading Job Ads Are Still Rising,” July 29, 2025. davron.net
  29. Clarify Capital employer survey (January 2025, n=1,000), via DAVRON. davron.net
  30. NOSSA, “Ghost Jobs in 2026,” January 9, 2026 (citing Resume Builder). nossahq.com
  31. Jobstrack.io, “Ghost Jobs Explained,” citing Jobright.ai (4.4M applications). jobstrack.io
  32. Columbia Law Review Forum, “Ghost Jobs,” November 2025. columbialawreview.org
  33. Congressional Research Service, “‘Ghost’ Job Postings,” IF12977, April 25, 2025. congress.gov
  34. Congressional Research Service, IF12977 (Ontario legislation). congress.gov
  35. PYMNTS, “Big Insurance Backs Away From AI Risk and Startups Rush In,” April 2026. pymnts.com
  36. PYMNTS (Munich Re standalone AI liability product, $2M–$50M coverage), April 2026. pymnts.com