Can employers require workers to use AI as a condition of employment?
Usually yes, but only until the AI mandate becomes a protected-trait, accommodation, state-law, or labor-law problem.
As of April 20, 2026, there is still no reported U.S. case saying an employer may, or may not, fire or refuse to hire someone solely because that person will not use AI tools. The nearest answer is older and more familiar: employment at will generally lets employers set methods of work, but that presumption gives way once the rule functions as an employment criterion and begins to interact with discrimination law, accommodation law, state AI-employment statutes, or labor law. The distinction that seems to matter most is between requiring output or proficiency in AI-shaped workflows and scoring employees on AI usage itself. The first looks more like ordinary managerial control. The second looks more like a selection or discipline criterion, which is where existing law starts to bite.
The default rule remains at-will employment. In most states, that means a private employer can usually change the tools, workflows, and productivity methods attached to a role. But at-will employment is a presumption, not a permission slip. California's Tameny v. Atlantic Richfield Co., 27 Cal. 3d 167 (1980), and New Jersey's Pierce v. Ortho Pharmaceutical Corp., 84 N.J. 58 (1980), are reminders that a discharge can still be actionable when it collides with a clear public policy. That matters here because a refusal to use AI will not always be about reluctance to adopt a new tool. It could instead be a refusal to put protected data into an unauthorized system, a refusal to rely on an inaccessible interface, or a refusal to carry out a discriminatory AI-scored workflow.
The firms are not saying mandatory AI use is clearly lawful. They are saying the question is governed by existing employment law, not by some new free-floating category called AI resistance. Littler's March 2025 piece treats AI policy as a way to narrow legal exposure, and it keeps returning to training, scope, and role design rather than to any theory that employers can simply declare AI use mandatory and be done with it. Ogletree's AI-policy podcast makes a similar point in plainer terms: an AI rule that exists only on paper is not enough, because the employer still needs a real purpose, enforceable guardrails, and some operational readiness behind the policy.
Fisher Phillips and Morgan Lewis are closer to each other than they first look. Fisher Phillips emphasizes policy, training, and continuity with ordinary labor and employment rules; its February 2026 testimony summary says employee rights under collective bargaining, wage-and-hour, and workplace-safety law still apply even when AI is involved. Morgan Lewis, meanwhile, keeps stressing that AI does not move responsibility away from the employer. Its 2024 pieces treat job descriptions, oversight, and anti-discrimination accountability as the real center of gravity, even when the system itself comes from a third party.
Sources for this answer
Commentary
A.1 NCSL, At-Will Employment - OverviewWhile the at-will employment doctrine serves as the default rule in the United States, it is subject to various common law and statutory exceptions that can limit an employer's ability to terminate an employee.
Employment relationships are presumed to be “at-will” in all U.S. states except Montana.
See NCSL, At-Will Employment - Overview.
Primary law
A.2 42 U.S.C. § 2000e-2(a)Title VII of the Civil Rights Act of 1964 prohibits employment discrimination based on race, color, religion, sex, or national origin and establishes the standard for proving such unlawful employment practices.
It shall be an unlawful employment practice for an employment agency to fail or refuse to refer for employment, or otherwise to discriminate against, any individual because of his race, color, religion, sex, or national origin
See 42 U.S.C. § 2000e-2(a).
Primary law
A.3 42 U.S.C. § 12112(b)(6)The Americans with Disabilities Act prohibits covered entities from discriminating against qualified individuals with disabilities in all aspects of employment and restricts the use of medical examinations and inquiries unless they are job-related and consistent with business necessity.
No covered entity shall discriminate against a qualified individual on the basis of disability in regard to job application procedures, the hiring, advancement, or discharge of employees, employee compensation, job training, and other terms, conditions, and privileges of employment.
See 42 U.S.C. § 12112(b)(6).
Primary law
A.4 Illinois Human Rights Act, 775 ILCS 5/2-102The Illinois Human Rights Act prohibits employers from engaging in various forms of discrimination, harassment, and unfair employment practices based on protected classes, including pregnancy, religion, and the use of artificial intelligence.
For an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes under this Article or to use zip codes as a proxy for protected classes under this Article.
See Illinois Human Rights Act, 775 ILCS 5/2-102.
Agency guidance
A.5 NLRB, General Counsel Issues Memo on Unlawful Electronic Surveillance and Automated Management PracticesThe NLRB General Counsel intends to challenge intrusive electronic surveillance and automated management practices that interfere with employees' Section 7 rights under the National Labor Relations Act.
General Counsel Jennifer Abruzzo announced her intention to protect employees, to the greatest extent possible, from intrusive or abusive electronic monitoring and automated management practices
See NLRB, General Counsel Issues Memo on Unlawful Electronic Surveillance and Automated Management Practices.
Commentary
A.6 Shopify CEO tells teams to consider using AI before growing headcountShopify has implemented a policy requiring teams to justify the need for additional headcount by demonstrating that the tasks cannot be performed by AI.
teams must demonstrate why they cannot get what they want done using AI
See Shopify CEO tells teams to consider using AI before growing headcount.
Law-firm commentary
A.7 Ogletree Deakins commentaryEmployers should implement comprehensive AI policies that include human oversight, independent verification of AI outputs, and proactive bias testing to mitigate legal and operational risks.
it is important that employers seek legal, ethical, and regulatory guidance when implementing AI platforms in their workspaces.
See Ogletree Deakins, The AI Workplace: A Guide on AI Policy Essentials.
Law-firm commentary
A.8 Morgan Lewis commentaryThe rapid adoption of artificial intelligence in the workplace has prompted significant federal and state regulatory oversight, including new compliance obligations for employers regarding bias audits, transparency, and risk management.
a poorly designed or trained AI tool has the potential to discriminate on a much larger scale.
See Morgan Lewis, AI in the Workplace: The New Legal Landscape Facing US Employers.
Case law
A.9 Tameny v. Atlantic Richfield Co., 27 Cal. 3d 167 (1980)An employee who is discharged for refusing to perform an illegal act at the direction of their employer may maintain a tort action for wrongful discharge because such a termination violates fundamental public policy.
when an employer's discharge of an employee violates fundamental principles of public policy, the discharged employee may maintain a tort action and recover damages traditionally available in such actions.
See Tameny v. Atlantic Richfield Co., 27 Cal. 3d 167 (1980).
Case law
A.10 Pierce v. Ortho Pharmaceutical Corp., 84 N.J. 58 (1980)New Jersey recognizes a common law cause of action for wrongful discharge where an employee is terminated for refusing to perform an act that violates a clear mandate of public policy, such as a statute, regulation, or professional code of ethics.
We hold that an employee has a cause of action for wrongful discharge when the discharge is contrary to a clear mandate of public policy.
See Pierce v. Ortho Pharmaceutical Corp., 84 N.J. 58 (1980).
Law-firm commentary
A.11 Littler Mendelson commentaryEmployers should implement tailored AI usage policies to mitigate legal and regulatory risks, ensure compliance, and maintain human oversight in workplace decision-making.
employers should consider adopting an AI policy to ensure that their use of AI is responsible, ethical, and legally compliant.
See Littler Mendelson, Considerations for Artificial Intelligence Policies in the Workplace.
Law-firm commentary
A.12 Fisher Phillips commentaryEmployers should implement a comprehensive workplace policy governing the use of generative AI to mitigate legal, security, and ethical risks associated with the technology.
A first step is developing a workplace GenAI policy.
See Fisher Phillips, The 10 Things All Employers Must Include in Any Workplace AI Policy.
Law-firm commentary
A.13 Fisher Phillips commentaryExisting employment and labor laws remain applicable to AI-driven workplace tools, necessitating that employers implement internal governance and transparent policies to manage risks and ensure compliance.
Walton cautioned them not to go too far, particularly since established employment and labor laws already apply.
See Fisher Phillips, What Responsible AI Use Means for Employers: 4 Takeaways from FP’s Recent Capitol Hill Testimony.
Law-firm commentary
A.14 Morgan Lewis commentaryThe OFCCP requires federal contractors to maintain full responsibility for EEO compliance and nondiscrimination obligations when utilizing AI systems, regardless of whether those systems are developed internally or provided by third-party vendors.
The OFCCP plans to treat AI models like other selection procedures, so federal contractors need to be able to articulate how their AI models operate and maintain records that substantiate these explanations.
See Morgan Lewis, US Department of Labor Publishes Guidelines Addressing Use of AI in Employment Decisions.
When does mandatory AI use create discrimination or accommodation risk?
The risk rises when AI use becomes a hiring, promotion, discipline, or performance criterion rather than a neutral work method.
Once the mandate becomes a hiring, promotion, or discipline criterion, Title VII is the first federal stop. The statute makes it unlawful to fail or refuse to hire or to discharge any individual because of a protected trait, and its disparate-impact provision asks whether the challenged practice is job related for the position in question and consistent with business necessity. An AI-use requirement could fit that framework if it disproportionately burdens a protected group through typing speed, language expectations, cognitive load, speech interaction, or constant-availability assumptions. The White House's April 23, 2025 executive order on disparate impact changed federal enforcement posture, but it did not amend the statute. That means the legal theory appears narrower as a policy priority than it was in 2024, yet it still exists as text and as private-claim architecture.
The ADA may be the cleaner fit. It bars discrimination against a qualified individual with a disability, requires reasonable accommodation absent undue hardship, and separately prohibits qualification standards or other criteria that screen out or tend to screen out an individual with a disability unless they are job-related and consistent with business necessity. That makes AI-use mandates unusually sensitive to interface design. If the worker can do the essential functions without personally using the mandated interface, the dispute could become whether AI use is actually essential or merely the employer's preferred method. The EEOC's AI materials do not speak directly to mandatory AI use, but they do say federal discrimination law applies to AI-enabled monitoring, productivity assessment, pay, promotion, and firing just as it applies to any other employment practice.
Religious accommodation is a narrower edge case, but it is no longer theoretical. Title VII defines religion broadly enough to trigger accommodation duties, and Groff v. DeJoy, 600 U.S. 447 (2023), now treats undue hardship as “substantial increased costs in relation to the conduct of its particular business”. Ogletree Deakins is already on record that employees may ask to avoid AI tasks, use manual methods, or avoid having AI usage counted in performance metrics. Those claims may remain rare. They are still part of the doctrinal map.
Ogletree is also the clearest firm source on religious objections. Its January 8, 2026 article expressly contemplates employees asking to avoid AI tasks, use manual alternatives, or avoid having AI abstinence count against them in performance review. That is important because it treats mandatory AI use as an accommodation problem before any court has done so.
Where firm commentary diverges, the disagreement is mostly about emphasis. Some firms focus on the executive branch's effort to cut back disparate-impact enforcement. Others emphasize that the underlying Title VII text remains untouched. Perhaps the cleanest synthesis is that the public-enforcement climate shifted, but the private-law question did not disappear.
It is also unsettled how often direct AI-use metrics can be defended as an essential function of a role rather than as a proxy for modernity, loyalty, or pace. Written job descriptions will likely matter, but perhaps not as much as the underlying reality of how the work gets done.
Another open question is how much the April 23, 2025 executive order will matter in private litigation. The firms agree that agency posture changed. They do not appear to agree that the statutory disparate-impact architecture went away.
Religious objections are plainly possible in theory and still hard to size in practice. Ogletree's article makes the doctrinal route visible, but there is still no reported line of cases about employees objecting to workplace AI use on that basis.
Sources for this answer
Primary law
B.1 42 U.S.C. § 2000e-2(a)Title VII of the Civil Rights Act of 1964 prohibits employment discrimination based on race, color, religion, sex, or national origin and establishes the standard for proving such unlawful employment practices.
It shall be an unlawful employment practice for an employment agency to fail or refuse to refer for employment, or otherwise to discriminate against, any individual because of his race, color, religion, sex, or national origin
See 42 U.S.C. § 2000e-2(a).
Primary law
B.2 White House, Restoring Equality of Opportunity and MeritocracyThis executive order establishes a federal policy to eliminate and deprioritize the enforcement of disparate-impact liability theories in civil rights and employment law, asserting that such theories are inconsistent with the Constitution.
It is the policy of the United States to eliminate the use of disparate-impact liability in all contexts to the maximum degree possible to avoid violating the Constitution, Federal civil rights laws, and basic American ideals.
See White House, Restoring Equality of Opportunity and Meritocracy.
Law-firm commentary
B.3 Mayer Brown commentaryPresident Trump's Executive Order directs federal agencies to deprioritize and seek to eliminate disparate-impact liability in enforcement actions and regulations, though its effect on private litigation remains uncertain due to existing statutory and judicial precedent.
The Order announces the Administration’s intent to “seek to eliminate the use of disparate-impact liability in all contexts to the maximum degree possible to avoid violating the Constitution, Federal civil rights laws, and basic American ideals.”
See Mayer Brown, Trump Executive Order Seeks to Eliminate Disparate-Impact Liability.
Law-firm commentary
B.4 Ogletree Deakins, President Trump Signs Executive Order Seeking to End Disparate Impact DiscriminationPresident Trump issued an executive order directing federal agencies to deprioritize the enforcement of disparate impact liability under statutes such as Title VII of the Civil Rights Act of 1964.
President Trump signed an executive order aimed at ending the legal theory of disparate impact discrimination by deprioritizing its enforcement within federal regulations, including Title VII of the Civil Rights Act of 1964.
See Ogletree Deakins, President Trump Signs Executive Order Seeking to End Disparate Impact Discrimination.
Primary law
B.5 42 U.S.C. § 12112(b)(6)The Americans with Disabilities Act prohibits covered entities from discriminating against qualified individuals with disabilities in all aspects of employment and restricts the use of medical examinations and inquiries unless they are job-related and consistent with business necessity.
No covered entity shall discriminate against a qualified individual on the basis of disability in regard to job application procedures, the hiring, advancement, or discharge of employees, employee compensation, job training, and other terms, conditions, and privileges of employment.
See 42 U.S.C. § 12112(b)(6).
Primary law
B.6 42 U.S.C. § 12111(8)A qualified individual under the ADA is one who can perform the essential functions of their job with or without reasonable accommodation, with the employer's written job description serving as evidence of those essential functions.
The term "qualified individual" means an individual who, with or without reasonable accommodation, can perform the essential functions of the employment position that such individual holds or desires.
See 42 U.S.C. § 12111(8).
Primary law
B.7 EEOC, What is the EEOC's role in AI?PDFThe EEOC maintains that federal anti-discrimination laws apply to the use of artificial intelligence in employment decisions, including instances where neutral practices result in an unjustifiable disparate impact.
These laws apply to the use of AI and other new technologies in employment just as they apply to other employment practices.
See EEOC, What is the EEOC's role in AI?.
Primary law
B.8 EEOC, Artificial Intelligence and the ADAThe EEOC provides guidance on how the Americans with Disabilities Act applies to the use of software, algorithms, and artificial intelligence in employment decision-making processes.
The Americans with Disabilities Act and the Use of Software, Algorithms, and Artificial Intelligence to Assess Job Applicants and Employees
See EEOC, Artificial Intelligence and the ADA.
Case law
B.10 Groff v. DeJoy, 600 U.S. 447 (2023)PDFSupports the cited proposition. (Groff v. DeJoy, 600 U.S. 447 (2023))
substantial increased costs in relation to the conduct of its particular business
See Groff v. DeJoy, 600 U.S. 447 (2023).
Primary law
B.9 42 U.S.C. § 2000e(j)The definition of religion under Title VII of the Civil Rights Act of 1964 was added to 42 U.S.C. § 2000e by the Equal Employment Opportunity Act of 1972.
Subsec. (j). Pub. L. 92–261, § 2(7), added subsec. (j).
See 42 U.S.C. § 2000e(j).
Law-firm commentary
B.11 Ogletree Deakins commentaryEmployers are legally required under Title VII and similar state laws to provide reasonable accommodations for an employee's sincerely held religious beliefs unless doing so imposes an undue hardship, defined by the Supreme Court as a substantial burden.
Federal and state laws require employers to provide reasonable accommodations for an employee’s religious beliefs, unless it would impose an undue hardship.
See Ogletree Deakins, The Mark of the Bot: When Employees Raise Religious Objections to Workplace AI Usage.
Which state AI employment laws apply to mandatory AI use?
State AI laws matter most when the mandate is tied to employment decisions such as hiring, discipline, discharge, or conditions of employment.
State AI law matters when the mandate is embedded in employment decision-making rather than ordinary workflow. Illinois now makes it unlawful to use AI in covered employment decisions if the use has a discriminatory effect, and it separately requires notice when AI is used for those purposes.recruitment, hiring, promotion, renewal of employment, selection for training or apprenticeship, discharge, discipline, tenure, or the terms, privileges, or conditions of employment California's automated decision system rules, effective October 1, 2025, do not create a standalone right to remain AI-free. They do make clearer that FEHA reaches AI or other computational processes used to make or assist employment decisions. Colorado is relevant but not yet operative here: the AI Act's effective date was pushed to June 30, 2026, so on April 20, 2026 it is background law, not current law.
Sources for this answer
Primary law
C.1 Illinois Human Rights Act, 775 ILCS 5/2-102The Illinois Human Rights Act prohibits employers from engaging in various forms of discrimination, harassment, and unfair employment practices based on protected classes, including pregnancy, religion, and the use of artificial intelligence.
For an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes under this Article or to use zip codes as a proxy for protected classes under this Article.
See Illinois Human Rights Act, 775 ILCS 5/2-102.
Primary law
C.2 Illinois Public Act 103-0804Illinois Public Act 103-0804 amends the Illinois Human Rights Act to prohibit employers from using artificial intelligence in employment decisions that results in discrimination or uses zip codes as a proxy for protected classes, and requires employers to provide notice when using such technology.
for an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes under this Article or to use zip codes as a proxy for protected classes under this Article.
See Illinois Public Act 103-0804.
Primary law
C.3 Illinois Department of Human Rights, Artificial Intelligence in Employment (P...Public Act 103-0804 mandates that employers utilizing artificial intelligence and automated decision-making systems in employment must maintain transparency and prevent discriminatory outcomes for protected classes.
This law establishes requirements for employers using artificial intelligence and automated decision-making systems in hiring and employment.
See Illinois Department of Human Rights, Artificial Intelligence in Employment (Public Act 103-0804).
Primary law
C.4 California Civil Rights Council Secures Approval for Regulations to Protect A...The California Civil Rights Council has adopted new regulations clarifying how existing state antidiscrimination laws apply to the use of artificial intelligence and automated-decision systems in employment.
The California Civil Rights Council today announced securing final approval for regulations to protect against potential employment discrimination as a result of the use of artificial intelligence, algorithms, and other automated-decision systems.
See California Civil Rights Council Secures Approval for Regulations to Protect Against Employment Discrimination Related to Artificial Intelligence.
Primary law
C.5 California Civil Rights Council Rulemaking ActionsThe California Civil Rights Council maintains a public record of its rulemaking actions, including the filing and effective dates of regulations approved by the Office of Administrative Law.
The Council’s Employment Regulations Regarding Automated Decision Systems have been approved by the Office of Administrative Law and were filed with the Secretary of State on June 27, 2025.
See California Civil Rights Council Rulemaking Actions.
Primary law
C.6 Colorado SB24-205Colorado SB24-205 establishes regulatory requirements for developers and deployers of high-risk artificial intelligence systems to prevent algorithmic discrimination and mandates transparency for consumer interactions with AI.
On and after February 1, 2026, the act requires a developer of a high-risk artificial intelligence system (high-risk system) to use reasonable care to protect consumers from any known or reasonably foreseeable risks of algorithmic discrimination in the high-risk system.
See Colorado SB24-205.
Primary law
C.7 Colorado SB25B-004Colorado SB25B-004 extends the effective date for compliance with the algorithmic transparency requirements originally established in Senate Bill 24-205 to June 30, 2026.
The act extends the effective date of the requirements of Senate Bill 24-205 to June 30, 2026.
See Colorado SB25B-004.
How should tech companies roll out mandatory AI use without creating metrics risk?
Separate AI-shaped output expectations from direct AI-usage scoring, and support the mandate with role design, training, and vendor controls.
The first consequence is that public corporate practice has already moved beyond voluntary experimentation. Shopify is the clearest example. Tobi Lütke's April 7, 2025 memo said “Reflexive AI usage is now a baseline expectation at Shopify” and tied AI use to headcount requests and review processes. Box's public posture is softer in language but similar in direction: an AI-first company, broader access to tools, company-wide upskilling, and changed output expectations. So the practical question is no longer whether companies will try this. It is how the legal system will describe what they are already doing.
The second consequence is that the line between a workflow rule and an employment criterion now carries most of the load. A company that says this role must meet output expectations inside an AI-shaped workflow looks closer to ordinary tool modernization. A company that says AI usage itself will be graded, and low usage will count against you looks closer to a new screening or discipline metric. Existing law is better developed for the second category than the first.
The third consequence is that training is becoming part of the liability story, not just part of change management. The Department of Labor's February and March 2026 AI-literacy materials do not create employer mandates, but they do make it easier to argue that AI competence can be a legitimate workforce-development goal. At the same time, a mandate with weak training makes the employer-side business necessity story less persuasive and the employee-side access story more persuasive. That is especially true where disability, language, or role design makes the interface itself part of the dispute rather than the output.
The fourth consequence is that enterprise vendor terms solve less than they first appear to solve. OpenAI, Microsoft, and Anthropic all say, in substance, that business-customer data is not used to train public models by default. That weakens the simple objection that a company mandate always forces employees to feed work product into public training. But it does not answer whether the tool is accessible, whether the rule has a disparate effect, whether a particular job really requires direct use, or whether the employer is grading the wrong thing.
The biggest unsettled question is whether courts will treat mandatory AI use as a simple proficiency requirement or as a new employment criterion in its own right. Perhaps the answer will turn less on the words in the policy than on what the employer actually measures.
Role-based mandates may cut both ways. They could help employers show job relatedness. They could also create selective-enforcement stories if the burdens and exceptions map awkwardly onto protected groups, older workers, disabled workers, or labor-protected employees. No court has yet drawn that line in the AI-use setting.
Sources for this answer
Commentary
D.1 Tobi Lutke, Reflexive AI usage is now a baseline expectation at ShopifySupports the cited proposition. (Tobi Lutke, Reflexive AI usage is now a baseline expectation at Shopify)
Reflexive AI usage is now a baseline expectation at Shopify
See Tobi Lutke, Reflexive AI usage is now a baseline expectation at Shopify.
Commentary
D.2 Shopify CEO tells teams to consider using AI before growing headcountShopify has implemented a policy requiring teams to justify the need for additional headcount by demonstrating that the tasks cannot be performed by AI.
teams must demonstrate why they cannot get what they want done using AI
See Shopify CEO tells teams to consider using AI before growing headcount.
Commentary
D.3 Aaron Levie, At Box, we are focused on building an AI-first companyBox has adopted a strategic framework for AI integration that prioritizes operational efficiency, reinvestment of cost savings into strategic growth, and robust data governance.
At Box, we are focused on building an AI-first company.
See Aaron Levie, At Box, we are focused on building an AI-first company.
Commentary
D.4 Box, AI-First Transformation: Box’s Principles, Strategy, and Execution FrameworkEffective enterprise AI transformation requires a governance model that combines centralized strategic guardrails with functional ownership to ensure scalable execution without creating bureaucratic bottlenecks.
Box’s AI executive sponsorship is led by Olivia Nottebohm, the company’s COO, Ravi Malick, the CIO, and Jessica Swank, Chief People Officer, to represent business, IT and people aspects — all of which are critical.
See Box, AI-First Transformation: Box’s Principles, Strategy, and Execution Framework.
Commentary
D.5 Box, Imagining the emerging role of the AI managerEffective AI governance requires the ongoing oversight of AI managers who are responsible for monitoring agent performance, managing change, and continuously recalculating the return on investment for AI tools.
Once an agent has moved through rollout and scaled adoption into the wild, the AI manager bears day-to-day responsibility for ensuring the agent is performing as expected, gathering feedback on how it’s being used, how it’s performing, and how the team can update and improve its functionality.
See Box, Imagining the emerging role of the AI manager.
Primary law
D.6 42 U.S.C. § 2000e-2(a)Title VII of the Civil Rights Act of 1964 prohibits employment discrimination based on race, color, religion, sex, or national origin and establishes the standard for proving such unlawful employment practices.
It shall be an unlawful employment practice for an employment agency to fail or refuse to refer for employment, or otherwise to discriminate against, any individual because of his race, color, religion, sex, or national origin
See 42 U.S.C. § 2000e-2(a).
Primary law
D.7 42 U.S.C. § 12112(b)(6)The Americans with Disabilities Act prohibits covered entities from discriminating against qualified individuals with disabilities in all aspects of employment and restricts the use of medical examinations and inquiries unless they are job-related and consistent with business necessity.
No covered entity shall discriminate against a qualified individual on the basis of disability in regard to job application procedures, the hiring, advancement, or discharge of employees, employee compensation, job training, and other terms, conditions, and privileges of employment.
See 42 U.S.C. § 12112(b)(6).
Primary law
D.8 Illinois Human Rights Act, 775 ILCS 5/2-102The Illinois Human Rights Act prohibits employers from engaging in various forms of discrimination, harassment, and unfair employment practices based on protected classes, including pregnancy, religion, and the use of artificial intelligence.
For an employer to use artificial intelligence that has the effect of subjecting employees to discrimination on the basis of protected classes under this Article or to use zip codes as a proxy for protected classes under this Article.
See Illinois Human Rights Act, 775 ILCS 5/2-102.
Agency guidance
D.9 U.S. Department of Labor releases AI literacy frameworkThe U.S. Department of Labor has issued an AI literacy framework designed to guide nationwide workforce development and skill-building efforts in alignment with the Trump Administration's AI strategy.
The U.S. Department of Labor’s Employment and Training Administration today published a framework for Artificial Intelligence literacy, providing a foundation to guide nationwide AI literacy efforts across workforce and education systems.
See U.S. Department of Labor releases AI literacy framework.
Primary law
D.10 U.S. Department of Labor, TEN 07-25The U.S. Department of Labor issued an Artificial Intelligence Literacy Framework to serve as a resource for program design and to promote AI literacy training within public workforce and education systems.
To issue the U.S. Department of Labor’s (DOL or Department) Artificial Intelligence (AI) Literacy Framework as a resource for program design and encourage expanded AI literacy training across the public workforce and education systems.
See U.S. Department of Labor, TEN 07-25.
Agency guidance
D.11 U.S. Department of Labor, Make America AI-ReadyThe U.S. Department of Labor launched the 'Make America AI-Ready' initiative to provide workers with foundational artificial intelligence skills and literacy through a structured, accessible training framework.
The initiative advances the Trump Administration’s commitment to equip American workers with foundational AI skills needed to succeed in an AI-driven economy
See U.S. Department of Labor, Make America AI-Ready.
Vendor documentation
D.12 OpenAI, Enterprise privacy at OpenAIOpenAI provides enterprise-level data privacy and security controls, including user ownership of inputs and outputs and a default policy against training models on customer data.
We do not train our models on your data by default
See OpenAI, Enterprise privacy at OpenAI.
Vendor documentation
D.13 Microsoft Learn, Data, Privacy, and Security for Microsoft 365 CopilotMicrosoft 365 Copilot maintains strict data privacy and security standards by prohibiting the use of organizational data for model training, enforcing existing user access permissions, and providing a copyright defense commitment for commercial customers.
Prompts, responses, and data accessed through Microsoft Graph aren't used to train foundation LLMs, including those used by Microsoft 365 Copilot.
See Microsoft Learn, Data, Privacy, and Security for Microsoft 365 Copilot.
Vendor documentation
D.14 Anthropic Privacy Center, Is my data used for model training?Anthropic does not use inputs or outputs from its commercial products to train its models by default, but may use data if a user explicitly provides feedback or opts into data usage.
By default, we will not use your inputs or outputs from our commercial products (e.g. Claude for Work, Anthropic API, Claude Gov, etc.) to train our models.
See Anthropic Privacy Center, Is my data used for model training?.
Law-firm commentary
D.15 Ogletree Deakins commentaryEmployers should implement comprehensive AI policies that include human oversight, independent verification of AI outputs, and proactive bias testing to mitigate legal and operational risks.
it is important that employers seek legal, ethical, and regulatory guidance when implementing AI platforms in their workspaces.
See Ogletree Deakins, The AI Workplace: A Guide on AI Policy Essentials.
Law-firm commentary
D.16 Morgan Lewis commentaryThe rapid adoption of artificial intelligence in the workplace has prompted significant federal and state regulatory oversight, including new compliance obligations for employers regarding bias audits, transparency, and risk management.
a poorly designed or trained AI tool has the potential to discriminate on a much larger scale.
See Morgan Lewis, AI in the Workplace: The New Legal Landscape Facing US Employers.
Law-firm commentary
D.17 Fisher Phillips commentaryExisting employment and labor laws remain applicable to AI-driven workplace tools, necessitating that employers implement internal governance and transparent policies to manage risks and ensure compliance.
Walton cautioned them not to go too far, particularly since established employment and labor laws already apply.
See Fisher Phillips, What Responsible AI Use Means for Employers: 4 Takeaways from FP’s Recent Capitol Hill Testimony.
When does mandatory AI use become a labor-law or monitoring problem?
The labor-law risk grows when the AI workflow also changes surveillance, quotas, pace-setting, discipline, or bargaining obligations.
Labor law sits beside all of this. The NLRB General Counsel's 2022 memo on electronic surveillance and automated management says these systems can unlawfully interfere with Section 7 rights, Stericycle tightened review of facially neutral work rules, and Endurance Environmental Solutions restored the clear and unmistakable waiver standard for unilateral changes affecting represented workers. So an AI-use mandate becomes a labor-law problem when it also changes surveillance, quotas, pace-setting, discipline, or bargaining obligations.
The fifth consequence is that unions and monitoring-heavy environments are different. In those settings, a mandatory AI workflow could be experienced less as a productivity tool and more as a combined system for pace, measurement, and discipline. That is where the NLRB materials matter most, and where an otherwise ordinary AI-use rule could start to look like a unilateral change in working conditions or a new surveillance regime.
Sources for this answer
Agency guidance
E.1 NLRB, General Counsel Issues Memo on Unlawful Electronic Surveillance and Automated Management PracticesThe NLRB General Counsel intends to challenge intrusive electronic surveillance and automated management practices that interfere with employees' Section 7 rights under the National Labor Relations Act.
General Counsel Jennifer Abruzzo announced her intention to protect employees, to the greatest extent possible, from intrusive or abusive electronic monitoring and automated management practices
See NLRB, General Counsel Issues Memo on Unlawful Electronic Surveillance and Automated Management Practices.
Agency guidance
E.2 NLRB, Board Adopts New Standard for Assessing Lawfulness of Work RulesThe NLRB established a new standard in Stericycle Inc. requiring employers to narrowly tailor work rules to advance legitimate business interests without chilling employees' Section 7 rights.
the NLRB issued a decision in Stericycle Inc., adopting a new legal standard for evaluating employer work rules challenged as facially unlawful under Section 8(a)(1) of the National Labor Relations Act.
See NLRB, Board Adopts New Standard for Assessing Lawfulness of Work Rules.
Agency guidance
E.3 NLRB, Board Returns to 'Clear and Unmistakable Waiver' StandardThe NLRB has returned to the 'clear and unmistakable waiver' standard for evaluating whether an employer may unilaterally change working conditions without bargaining with a union.
restored the “clear and unmistakable” waiver standard for evaluating employers’ contractual defenses to allegations that they have unlawfully changed the working conditions of union-represented employees
See NLRB, Board Returns to 'Clear and Unmistakable Waiver' Standard.