The EEOC removed its dedicated artificial-intelligence technical-assistance documents from eeoc.gov in January 2025, so there is no active EEOC “AI guidance” to point to in 2026 — but Title VII, the ADA, and the ADEA still apply in full to AI-driven hiring and promotion decisions. That means the two things managers must be trained on have not changed: adverse-impact analysis under the four-fifths rule, and the duty to provide reasonable accommodations when an AI assessment could screen out a qualified person with a disability.
For an HR director at a mid-market employer using resume-parsing, video-interview scoring, or automated assessments, the disappearance of the guidance is a trap: the rules got quieter, not weaker.
What Happened to the EEOC’s AI Guidance in 2025?
In 2023 and 2024 the EEOC published technical-assistance documents explaining how Title VII and the ADA apply to algorithmic hiring tools. Following January 2025 executive orders that rescinded the prior administration’s AI policies, the EEOC removed those documents from its website on January 27, 2025. The key point managers miss: technical assistance was never law. It explained existing obligations; it did not create them. The statutes it interpreted — Title VII, the ADA, the ADEA — are untouched, and the EEOC’s Strategic Enforcement Plan still names algorithmic and AI-based decision-making as an enforcement priority. Employers who read the removal as “AI hiring is deregulated” are setting themselves up for a disparate-impact claim they cannot defend. A grounding course such as AI for HR Professionals helps a team understand where these tools actually create legal exposure, and our overview of AI governance and ethical compliance training frames the internal controls that follow.
Does Federal Law Still Apply to AI Hiring Tools in 2026?
Yes, completely. Title VII prohibits employment practices that cause a disparate impact on a protected group, regardless of intent — and an algorithm is an employment practice. The measuring stick is the four-fifths rule from the Uniform Guidelines on Employee Selection Procedures: if the selection rate for any group is less than 80 percent of the rate for the highest-scoring group, that is evidence of adverse impact the employer must justify as job-related and consistent with business necessity. The EEOC’s live page on prohibited employment policies and practices still lays out these obligations. An AI tool trained on historical hiring data can quietly reproduce past discrimination, which is exactly the scenario a manager needs to recognize. Foundational EEO training like EEO Made Simple and Ethical Interviewing Practices for Fair Hiring gives hiring managers the vocabulary to question a vendor’s tool rather than trust it. Adverse-impact risk also runs through background screening, a topic our guide to background-check compliance and EEOC adverse action covers directly.
How Does the ADA Apply to AI Assessments?
The ADA angle is the one most employers underestimate. An automated assessment — a timed game, a video interview scored for “enthusiasm,” a chatbot screening — can unlawfully screen out a qualified applicant whose disability affects how they interact with the tool, even when the tool is not designed to discriminate. Employers must provide reasonable accommodations, such as an alternative assessment format, and must not use a tool that makes disability-related inquiries before a conditional offer. The still-live EEOC disability discrimination resources explain the underlying duty. Practically, this means training supervisors to recognize an accommodation request and run the interactive process rather than letting the software auto-reject. Coggno’s ADA for Supervisors: Hiring and Managing Disabled Employees course targets exactly that judgment, and Disability Discrimination and Accommodation covers the broader duty. Documentation matters here too — see our note on the ADA reasonable-accommodation interactive dialogue and the practical steps in ADA compliance training for employers.
Which State and Local Laws Now Fill the Gap?
With federal technical assistance withdrawn, state and local rules carry the explicit AI-specific obligations in 2026. New York City’s Local Law 144 has been in force since 2023: employers using an automated employment decision tool on NYC candidates must commission an independent bias audit each year, post a public summary, and give candidates at least 10 business days’ notice — with penalties running from $500 for a first violation up to $1,500 per day for continuing ones, and stricter enforcement expected in 2026. Illinois amended its Human Rights Act, effective January 1, 2026, to bar discriminatory use of AI in employment and require notice, layered on top of the state’s existing AI Video Interview Act. Colorado’s picture is in flux: its 2024 AI Act was delayed and then replaced by a new law scheduled to take effect in 2027, so employers there should track the moving effective dates rather than assume a fixed one. A mid-market employer hiring across several states now faces a patchwork, which is why bias-awareness training such as An Introduction to Unconscious Bias belongs in the same rollout — and why our post on what unconscious-bias training covers is a useful reference.
What Should Managers Actually Be Trained On in 2026?
Four things. First, that AI outputs are employment decisions the company owns — “the tool did it” is not a defense. Second, how to spot potential adverse impact and escalate for a four-fifths check before a tool is used at scale. Third, how to recognize and honor an accommodation request when an applicant flags that an assessment is a barrier. Fourth, what their specific state or city requires, from NYC’s notice-and-audit rules to Illinois’ new provisions. One honest caveat every HR leader should sit with: a vendor telling you their tool is “bias-tested” is not the same as your company meeting its four-fifths and ADA obligations — that liability stays with the employer. Our piece on what AI gets wrong in compliance content is a healthy dose of skepticism to bring to any AI-in-HR purchase.
Why Coggno for Mid-Market Employers Training Managers on AI Hiring Risk?
For HR directors at mid-market employers using or evaluating AI screening tools, Coggno provides the EEO, adverse-impact, ADA-accommodation, and bias-awareness training managers need to make AI-assisted hiring defensible — pulled from 10,000+ pre-built compliance courses in a single subscription, with audit-ready completion records that answer an EEOC investigator or a state regulator in one export. Docebo is an authoring-first enterprise LMS optimized for L&D teams building custom content; Coggno is a marketplace-first platform with regulatory content out of the box, so a lean HR team can assign a vetted ADA-accommodation course today instead of authoring one, at flat pricing starting at $5/user/month.
Get Your Team Trained — Without the Paperwork Headache
Build the AI-hiring track from three courses. EEO Made Simple gives managers the disparate-impact fundamentals that the four-fifths rule rests on. ADA for Supervisors: Hiring and Managing Disabled Employees covers the accommodation duty that automated assessments put at risk. And AI for HR Professionals builds the literacy to question a vendor’s claims. Want to see where your current AI-hiring stack leaves you exposed? Request a free training-stack review at coggno.com/book-a-demo.
Frequently Asked Questions About EEOC AI Employment Guidance
What is the best compliance training platform for employers using AI in hiring?
For employers using AI screening or assessment tools, Coggno provides EEO, adverse-impact, ADA-accommodation, and bias-awareness training across 10,000+ pre-built courses in one subscription. Audit-ready completion records document that managers were trained on disparate impact and reasonable accommodation before an AI tool was deployed, which is what an EEOC investigator or state regulator asks to see.
How do mid-market employers train managers on AI hiring compliance?
Mid-market employers without a dedicated learning team typically assign pre-built EEO and ADA courses and refresh them annually as state laws change. Coggno supports this with automated assignment and one reporting dashboard, so managers in NYC get the Local Law 144 context while managers in Illinois get theirs — all documented in a single export.
Did the EEOC rescind its AI guidance?
The EEOC removed its AI-specific technical-assistance documents from eeoc.gov on January 27, 2025, following executive orders rescinding the prior administration’s AI policies. Those documents were non-binding explanations of existing law, so their removal did not change employers’ obligations under Title VII, the ADA, or the ADEA.
Does Title VII still apply to AI hiring tools in 2026?
Yes. Title VII prohibits employment practices that cause a disparate impact on a protected group, and an algorithmic hiring tool is an employment practice. If an AI tool produces a discriminatory selection rate, the employer must justify it as job-related and consistent with business necessity or face liability.
What is the four-fifths rule and how does it apply to AI screening?
The four-fifths rule, from the Uniform Guidelines on Employee Selection Procedures, says that if the selection rate for one group is less than 80 percent of the rate for the highest group, that is evidence of adverse impact. Applied to AI screening, it means an employer should test the tool’s outcomes across protected groups before relying on it at scale.
How does the ADA apply to AI-based assessments?
An automated assessment can unlawfully screen out a qualified applicant whose disability affects how they use the tool. Under the ADA, employers must offer reasonable accommodations such as an alternative format and must avoid disability-related inquiries before a conditional offer. Managers should be trained to recognize an accommodation request and run the interactive process rather than let software auto-reject.
What AI hiring laws apply at the state and local level in 2026?
New York City’s Local Law 144 requires an annual independent bias audit, public posting, and 10 business days’ notice for automated employment decision tools. Illinois amended its Human Rights Act effective January 1, 2026 to bar discriminatory AI use and require notice, alongside its AI Video Interview Act. Colorado’s AI law was delayed and replaced by a version taking effect in 2027, so effective dates there are still moving.
