Department of Labor

Authors’ note: For this report, the authors use the definition of artificial intelligence (AI) from the 2020 National Defense Authorization Act, which established the National Artificial Intelligence Initiative.¹ This definition was also used by the 2023 “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.”² Similarly, this report makes repeated reference to “Appendix I: Purposes for Which AI is Presumed to be Safety-Impacting and Rights-Impacting” of the 2024 OMB M-24-10 memo, “Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence.”³

The U.S. Department of Labor (DOL) oversees numerous statutes, from the Fair Labor Standards Act (FLSA) to the Family and Medical Leave Act (FMLA), that can potentially help address the challenges and opportunities of artificial intelligence (AI) as it affects workers. Governing for Impact (GFI) and the Center for American Progress have extensively researched these existing authorities in consultation with numerous subject matter experts. However, the goal is to provoke a generative discussion about the following proposals, rather than outline a definitive executive action agenda. Each potential recommendation will require further vetting before agencies act. Even if additional AI legislation is needed, this menu of potential recommendations to address AI demonstrates that there are more options for agencies to explore beyond their current work and that agencies should immediately utilize existing authorities to address AI.

The proliferation of AI and automated algorithmic technologies poses both macro and micro challenges for workers. At one extreme, sufficiently advanced AI may displace entire occupation categories, putting thousands or millions of Americans out of work. But such dramatic predictions can also overshadow how AI and automated technologies already play a role in shifting worker power to employers and denying workers statutory protections.

While the ultimate scale of workplace disruption remains unknown, the DOL is responsible for implementation and enforcement of several statutes that protect and empower workers.⁴ As President Joe Biden acknowledged in his 2023 “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,” automated systems and other technology deployed in the workplace hold great potential to improve working conditions.⁵ But they can also pose grave risks to workers’ rights and safety if not used carefully or implemented without worker input.⁶ This section explains some of the known risks of AI in the workplace and identifies DOL-enforced statutes that could be used to address them through regulations, subregulatory guidance, and enforcement practices. Among other authorities, the DOL could use these statutes to ameliorate known harms by updating wage and hour regulations, guarding workers’ safety and health against the negative impacts of automated management, and ensuring that automated benefits administration is transparent and fair.

AI may cause harm to American workers in several known and unknown ways. Although certainly not exhaustive, the known risks can be roughly grouped into seven categories:

• Discrimination: Algorithmic bias refers to an algorithm’s tendency toward replicating or amplifying human biases due to unrepresentative or incomplete training data or reliance on information that reflects historical inequalities.⁷ As one of the eight primary policies and principles of the Biden administration’s approach to AI, the 2023 executive order on AI noted: “From hiring to housing to healthcare, we have seen what happens when AI use deepens discrimination and bias, rather than improving quality of life. Artificial Intelligence systems deployed irresponsibly have reproduced and intensified existing inequities, caused new types of harmful discrimination, and exacerbated online and physical harms.”⁸

The May 2024 DOL “Artificial Intelligence and Worker Well-being: Principles for Developers and Employers” lists as a priority that “AI systems should not violate or undermine workers’ … anti-discrimination and anti-retaliation protections.”⁹ As the Equal Employment Opportunity Commission has identified,¹⁰ algorithmic bias can be embedded in technologies that employers increasingly use to make hiring and retention decisions. Such bias can surreptitiously disadvantage workers or applicants based on any number of protected characteristics by shaping recruitment efforts toward “relevant” job seekers and narrowing the candidate pool through automated interview technology or based on historical employment data.¹¹ For example, there is ample evidence that AI-driven interview software, which films interview responses and assesses candidates’ performance, is “definitionally discriminatory”¹² against individuals with disabilities.¹³

• Access to benefits: An increased use of AI in evaluating claims for benefits such as health insurance and unemployment compensation or making investment decisions in employer-sponsored retirement accounts could pose significant risk to workers’ well-being.¹⁴ This could occur because, for example, an algorithmic system denies claims at a higher rate than a human,¹⁵ or because an AI-enabled investment allocation technology could prioritize investments that trigger payment of transaction fees and commissions to a brokerage over investments that would maximize an employee’s retirement savings.¹⁶
• Safety and health: As GFI has documented in its prior regulatory advocacy work, electronic surveillance and automated management (ESAM) can pose risks to workers’ physical and mental safety and health.¹⁷ Employers’ unrestrained use of ESAM can result in an unsustainable pace of work that increases accident rates and musculoskeletal strain.¹⁸ The Office of Management and Budget (OMB) M-24-10 memorandum on “Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence” specifically identified AI used to “control or significantly influence the outcomes of … physical movements of robots or robotic appendages within a workplace, school, housing, transportation, medical, or law enforcement setting” as presumed to be safety-impacting.¹⁹ ESAM can also heighten mental health stress for workers as they labor under extreme pressure with little privacy.²⁰ As ESAM technologies become more sophisticated, they create even more risks to mental health—for instance, by increasing pressure on workers via technology that detects and measures emotions and thoughts. This particular risk was cited as an AI purpose presumed to be rights-impacting by the OMB M-24-10 AI memo.²¹ The May 2024 DOL AI principles prioritize that “AI systems should not violate or undermine workers’ … health and safety rights.”²²
• Wage and hour compliance: AI and the remote, on-demand work that it can enable raise thorny questions about employers’ obligations under wage and hour laws. As technology blurs the line between work and nonwork time, it may become more difficult to assess what time is compensable, and it therefore should be taken into account in assessing compliance with minimum wage and overtime laws.²³ Additionally, if AI takes over tasks that involve discretion, creativity, and supervision, or if automated timekeeping software automatically reduces wages for time spent off-task, workers who previously were not eligible for overtime compensation may become eligible.²⁴ Other risks include opacity and manipulation in algorithmic wage-setting technologies²⁵ and digital wage theft enabled by an outdated regulatory regime.²⁶ The May 2024 DOL AI principles highlight as a priority that “AI systems should not violate or undermine workers’ … wage and hour rights.”²⁷
• Misclassification: Most federal laws that set labor standards apply only to employees, rather than independent contractors.²⁸ Whether a worker is an employee or an independent contractor, as well as whether an indirect employer—for example, a parent company—counts as an employer for statutory purposes, is largely determined based on the amount of control an entity possesses over the worker.²⁹ AI-enabled, always-on ESAM makes it easier for companies to exert control over workers while avoiding the traditional hallmarks of employer control, such as on-site, human supervision.³⁰ If companies are able to avoid being classified as a joint employer of a worker or retain independent contractor status for their workers by supplanting traditional modes of control with virtual control enabled by ESAM, they may skirt their obligations under a host of employment laws.³¹ For example, McDonald’s has long argued that it is not the joint employer of employees in franchised stores despite the tight control that headquarters exerts over franchise employees by tracking their productivity through point-of-sale technology.³²
• Worker power and datafication: AI can disempower workers by disrupting organizing efforts³³ through, for example, surveillance and scheduling tricks and by accelerating “worker datafication.” This refers to employers’ ravenous collection, use, and resale of workers’ data without regard for workers’ ownership of, privacy regarding, or ability to benefit from the data.³⁴ The May 2024 DOL AI principles prioritize that “AI systems should not violate or undermine workers’ right to organize” and that “Workers’ data collected, used, or created by AI systems should be limited in scope and location, used only to support legitimate business aims, and protected and handled responsibly.”³⁵
• Workforce training and displacement: When people think about AI’s impact on workers, a common first thought is the potential for mass layoffs and job displacement. According to one estimate, activities that account for up to 30 percent of the hours currently worked across the U.S. economy could be automated.³⁶ Of course, the automation of existing jobs is only part of the story, as experts expect AI to create a new wave of jobs associated with the technological revolution, for which American workers must be prepared. The World Economic Forum estimated in 2020 that while AI would displace 85 million jobs worldwide by 2025, the technology would also create 97 million new roles.³⁷ However, even if the net impact on jobs is positive, there is still the potential for significant point-in-time job losses or net losses in particular geographies, possibly at a more rapid pace than the United States has experienced before.³⁸ The DOL AI principles note the “risks that workers will be displaced entirely from their jobs by AI” and highlight “Supporting Workers Impacted by AI” as a principle, saying that “[e]mployers should support or upskill workers during job transitions related to AI.”³⁹

The application of AI in the realm of employment law is not all bad for workers. The technology holds great promise for federal agencies, including the DOL, to augment their enforcement capabilities. For instance, it could be used to analyze reams of wage and hour data, trace patterns, and help identify employers for further investigation of potential statutory violations—for example, failure to report expenditures on union avoidance consultants—and to ensure payment of prevailing, minimum, or overtime wages. In particular, the use of AI for analyzing, investigating, and auditing prevailing wage enforcement represents an early opportunity, given the existing data on wages and benefits. Rapid data collection and analysis fueled by AI could help occupational safety and health experts draw conclusions about workplace characteristics and job conditions that are most likely to lead to injury and illness.⁴⁰ New research highlights how AI-aided enforcement strategies could dramatically reduce workplace injuries.⁴¹ AI could also be used to spotlight further review instances where investment decisions and employee benefit determinations may be rigged against workers.

The DOL has already taken promising action on AI and plans to take more. For example, it published a blog post explaining what the White House’s AI Bill of Rights means for workers.⁴² The Office of Federal Contract Compliance Programs (OFCCP) collects information from federal contractors about their use of AI in recruitment, screening, and hiring.⁴³ The Office of Disability Employment Policy funds various projects related to AI in employment, such as the Partnership on Employment and Accessible Technology’s “AI & Disability Inclusion Toolkit.”⁴⁴

The 2023 executive order on AI also directed the DOL to take several steps, including preparing reports for the president and publishing guidance about wage and hour and health and safety risks related to AI.⁴⁵ In response, the DOL issued a field assistance bulletin in April 2024 that describes how various federal labor standards apply to employers who use AI to manage their workforces.⁴⁶ For example, the guidance document addresses how AI-enabled employee monitoring tools that track keystrokes and other activities could unlawfully deprive workers of compensation for working time spent on noncomputer tasks.⁴⁷ Additionally, the bulletin highlights the potential danger of embedding errors in automated employment tools because of the potential to affect a large group of workers quickly.⁴⁸

In response to the 2023 executive order on AI, the Department of Labor’ published “Artificial Intelligence and Worker Well-being: Principles for Developers and Employers” in May 2024.⁴⁹ These eight principles include a North Star of “Centering Worker Empowerment,” along with priorities that include ethical AI development, transparency, and protection of labor and employment rights.⁵⁰

The executive order also directed the DOL to issue guidance for “[f]ederal contractors regarding nondiscrimination in hiring involving AI and other technology-based hiring systems,”⁵¹ which the DOL released in May 2024 as the “Artificial Intelligence and Equal Employment Opportunity for Federal Contractors.”⁵² The guidance includes a discussion of the types of bias that can be embedded in algorithmic decision-making processes and explains how federal contractors are responsible for compliance with nondiscrimination statutes regardless of whether their hiring decisions involve automation.⁵³ Additionally, the guidance explicitly states that federal contractors cannot delegate compliance responsibilities to outside entities—including vendors—and provides several promising practices to maintain compliance.⁵⁴

Finally, the OMB M-24-10 AI memo—primarily applicable to agencies’ procurement of AI software—focused on AI use cases that are rights-impacting, including:

… [d]etermining the terms or conditions of employment, including pre-employment screening, reasonable accommodation, pay or promotion, performance management, hiring or termination, or recommending disciplinary action; performing time-on-task tracking; or conducting workplace surveillance or automated personnel management.⁵⁵

The jurisdiction of the DOL in enforcing an array of statutes is limited, at least in part, by the department’s interpretation of the statutes’ definitions of “employee” and “employer.”⁵⁶ By the DOL’s terms, the laws typically do not protect individuals working as independent contractors and instead only apply to employees. Additionally, an employer can only be held accountable concerning the statutory rights of an employee if they are found to be in an employment relationship with a particular worker.⁵⁷ This distinction comes into play particularly when a lead firm contracts parts of its workforce to subcontractors or franchisees. While statutes vary, a key factor in determining whether a worker is an independent contractor or an employee and whether a lead firm is a joint employer of a subcontractor’s employee is the amount of control that the employer exercises or has the authority to exercise over the worker.⁵⁸

The preamble to the DOL’s rule defining “employee” under the FLSA, and therefore also the FMLA,⁵⁹ recognizes this connection. It explicitly states that “whether the employer uses technological means of supervision (such as by means of a device or electronically)” is a “[fact] relevant to the employer’s control over the worker.”⁶⁰ The preamble to the rule also discussed the role that electronic monitoring plays in a control analysis.⁶¹ The DOL could also consider explicitly recognizing the use of ESAM as an indicator of control for joint employer recognition under employment statutes.⁶² As GFI explained in a comment to the National Labor Relations Board (NLRB) in 2023, surveillance practices allow lead firms to tightly control their subcontractors’ or franchisees’ employees with whom the companies disclaim having an employment relationship.⁶³ Ensuring that control exerted and reserved by ESAM systems is considered in joint employer analyses, even in the absence of traditional hallmarks of employer control—such as on-site, real-time, human supervision—will ensure workers can hold entities that control them accountable for the entities’ obligations under employment law.

This section explains how some statutes currently enforced by the DOL could apply to AI. As explained in the introduction to this report, this list is by no means exhaustive, and each potential proposal would benefit from additional research and vetting.

At 29 U.S.C. § 211(c), the FLSA requires employers to “make, keep, and preserve” records of “wages, hours, and other conditions and practices of employment” and submit reports to the Wage and Hour Division (WHD) administrator “as he shall prescribe by regulation … as necessary or appropriate for the enforcement” of the statute.⁶⁴ The DOL has used this authority to issue regulations at 29 C.F.R. Part 516 that, among other things, require employers to keep records that include regular hourly rates of pay, records of retroactive payments of wages, and documentation demonstrating whether the employee qualifies for any exemptions under the FLSA.⁶⁵ 

At 29 U.S.C. § 211(a), the FLSA states:

The [WHD] Administrator or his designated representatives may investigate and gather data regarding the wages, hours, and other conditions and practices of employment in any industry subject to this chapter, and may enter and inspect such places and such records (and make such transcriptions thereof), question such employees, and investigate such facts, conditions, practices, or matters as he may deem necessary or appropriate to determine whether any person has violated any provision of [the FLSA], or which may aid in the enforcement of the provisions of [the FLSA.]⁶⁶

At 29 U.S.C. § 206(a), the FLSA requires most employers to pay most employees a minimum wage for all hours worked.⁷³ At 29 U.S.C. § 207(a)(1), the FLSA requires most employers to pay most employees 1.5 times their regular rate of pay for all hours worked in excess of 40 hours per week.⁷⁴ At 29 U.S.C. § 254(a), the Portal-to-Portal Act amended the FLSA to exempt from “hours worked” time spent commuting and time spent on “activities which are preliminary to or postliminary to” an employee’s principal activities of work.⁷⁵

In addition to the statutory amendments’ attempted clarification of what time is considered compensable for minimum wage and overtime purposes, the WHD has issued several interpretive regulations, organized at 29 C.F.R. Part 785, and pieces of guidance, field assistance bulletins,⁷⁶ on the subject. At 29 C.F.R. § 781.11–13, for example, the WHD explains that an employee’s time is compensable if the employer knows or has reason to know that the employee is engaged in work, and that principle applies to work completed away from the job site “or even at home.”⁷⁷ Citing administrative ease, at 29 C.F.R. § 785.47, the regulations draw on judicial precedent to set forth the WHD’s de minimis rule, which exempts “insubstantial or insignificant periods of time beyond the scheduled working hours” from compensability.⁷⁸ Similarly, based on administrability rationales, current regulations permit employers to round timesheets to the nearest quarter-hour.⁷⁹

Besides bona fide meal breaks,⁸⁰ which are not compensable, most time in a day between an employee’s first performance of a “principal” activity and when the employee ceases such activity is compensable.⁸¹ Activities that are “principal” are those that a worker is “employed to perform,” rather than those that are preliminary or postliminary, such as commuting.⁸²

The unemployment compensation system is a joint federal-state scheme that provides support through individual benefit payments.⁸⁶ The federal statute establishes broad requirements for the program, but the specifics are determined by state laws, which are administered with DOL oversight.⁸⁷

At 42 U.S.C. § 503(a)(1), the federal unemployment compensation statute requires states implementing laws to include “[s]uch methods of administration … as are found by the Secretary of Labor to be reasonably calculated to insure full payment of unemployment compensation when due.”⁸⁸ If the DOL determines that a state violates §503(a)(1), the statute directs the DOL, at 42 U.S.C. § 503(b), to stop payment to the state unemployment agency.⁸⁹ To ensure compliance, at 42 U.S.C. § 503(a)(6), the statute requires states to make reports “in such form and containing such information, as the Secretary of Labor may from time to time require, and compliance with such provisions as the Secretary of Labor may from time to time find necessary to assure the correctness and verification of such reports.”⁹⁰ 42 U.S.C. § 1302 directs the secretary of labor to “make and publish such rules and regulations … as may be necessary to the efficient administration of” several social welfare programs, including unemployment compensation.⁹¹

The 2023 executive order on AI specifically directs the secretary of labor to assess how unemployment insurance “could be used to respond to possible future AI-related disruptions.”⁹² The OMB M-24-10 AI memo specifically declares AI used for:

Making decisions regarding access to, eligibility for, or revocation of critical government resources or services; allowing or denying access—through biometrics or other means (e.g., signature matching)—to IT systems for accessing services for benefits; detecting fraudulent use or attempted use of government services; assigning penalties in the context of government benefits [to be presumptively rights-impacting and subject to specific minimum risk management practices.]⁹³

Congress enacted the Occupational Safety and Health (OSH) Act with the purpose of “[assuring] so far as possible every working man and woman in the Nation safe and healthful working conditions.”¹⁰⁰ An examination of the congressional record makes clear that Congress established the Occupational Safety and Health Administration (OSHA) because the field of occupational safety and health was changing quickly.¹⁰¹ Congress decided that it needed to empower a federal agency with the authority to keep up with changes in the organization of work and establish rules to protect workers. As a congressional report explained, “technological advances and new processes in American industry have brought numerous new hazards to the workplace.”¹⁰² New “processes are being introduced into industry at a much faster rate than the present meager resources of occupational health can keep up with.”¹⁰³

At 29 U.S.C. § 654(a)(2), the OSH Act requires employers to comply with “occupational safety and health standards promulgated under” the law.¹⁰⁴ At 29 U.S.C. § 652(8), the statute defines “occupational safety and health standard” as “a standard which requires conditions, or the adoption or use of one or more practices, means, methods, operations, or processes, reasonably necessary or appropriate to provide safe or healthful employment and places of employment.”¹⁰⁵ At 29 U.S.C. § 655(b), the statute establishes how OSHA may prescribe such standards.¹⁰⁶ OSHA has used this authority repeatedly to issue regulations at 29 C.F.R. Part 1910 that protect workers from workplace exposure to lead, ergonomic risk,¹⁰⁷ and many other hazards. OSHA also has experience regulating and issuing guidance about mental health hazards that threaten workers.¹⁰⁸

In addition to the law’s substantive requirements, the OSH Act requires employers to record and report information, including about workplace injuries, to OSHA. At 29 U.S.C. § 657(c)(1), the statute mandates that employers “make, keep and preserve, and make available to the Secretary [of Labor] … such records regarding his activities relating to this Act as the Secretary . . . may prescribe by regulation as necessary or appropriate for the enforcement of this Act or for developing information regarding the causes and prevention of occupational accidents and illnesses.”¹⁰⁹ At 29 U.S.C. § 657(c)(2), the statute directs the secretary of labor to prescribe regulations “requiring employers to maintain accurate records of, and to make periodic reports on, work-related deaths, injuries and illnesses other than minor injuries.”¹¹⁰

At 29 U.S.C. § 671(d), the statute directs the National Institute for Occupational Safety and Health (NIOSH) to conduct, upon its own initiative or “upon the request of the Secretary” of labor or health and human services, research “necessary for the development of criteria for new and improved occupational safety and health standards.”¹¹¹

Congress enacted the Employee Retirement Income Security Act (ERISA) to establish a comprehensive regulatory scheme for employee pension and welfare benefit plans, including group health insurance plans, offered by private sector employers.¹²¹ The act creates protections for plan participants and beneficiaries by setting requirements related to disclosure and reporting about decisions regarding benefit eligibility, benefit accrual, investing and plan administration, and plan funding.¹²²

At 29 U.S.C. § 1133, the statute requires:

In accordance with regulations of the Secretary [of Labor], every employee benefit plan shall—(1) provide adequate notice in writing to any participant or beneficiary whose claim for benefits under the plan has been denied, setting forth the specific reasons for such denial, written in a manner calculated to be understood by the participant, and (2) afford a reasonable opportunity to any participant whose claim for benefits has been denied for a full and fair review by the appropriate named fiduciary of the decision denying the claim.¹²³

At 29 U.S.C. § 1022, the statute requires that a “summary plan description of any employee benefit plan shall be furnished to participants and beneficiaries,” which shall include a description of the “circumstances which may result in disqualification, ineligibility, or denial or loss of benefits.”¹²⁴ At 29 U.S.C. § 1029(c), the statute authorizes the secretary of labor to “prescribe the format and content of the summary plan description.”¹²⁵

At 29 U.S.C. § 1135, the statute permits the secretary of labor to “prescribe such regulations as he finds necessary or appropriate to carry out” ERISA’s requirements.

At 29 U.S.C. § 1104, ERISA imposes responsibilities on plan fiduciaries, who are individuals that are responsible for plan management and operations.¹³¹ Among them are the duties of prudence and loyalty.¹³² At 29 U.S.C. § 1104(c)(5), the statute requires default investment allocations for retirement savings plans to be “invested by the plan in accordance with regulations prescribed by the Secretary.”¹³³ The DOL recently issued a proposed rule to revise the scope of ERISA’s coverage of investment advice fiduciaries to keep pace with the modern economy.¹³⁴

Congress passed the Labor Management Reporting and Disclosure Act (LMRDA) to level the playing field between management and labor by providing transparency for workers, government, and advocates into the complex anti-union persuasion industry.¹³⁸ The law requires a series of disclosures from unions, employers, and union-avoidance consultants and law firms to ensure that workers know the sources of the huge sums of money that go into urging them one way or another on unionization.¹³⁹

At 29 U.S.C. § 433(a)(3), the LMRDA requires employers to file a report to DOL “in a form prescribed by” the secretary of labor if the employer makes “any expenditure, during the fiscal year, where an object thereof, directly or indirectly, is to interfere with, restrain, or coerce employees in the exercise of the right to organize and bargain collectively through representatives of their own choosing, or is to obtain information concerning the activities of employees or a labor organization in connection with a labor dispute involving such employer.”¹⁴⁰ The DOL has used this authority to specify what information employers must report in LM-10 forms, including, most recently, the DOL’s proposed rule requiring employers to identify themselves as federal contractors.¹⁴¹Additionally, the LM-10 form instructions identify what types of transactions employers must report.¹⁴²

Congress enacted the Worker Adjustment and Retraining Notification (WARN) Act to help workers and communities prepare for economic dislocation caused by mass job losses.¹⁴⁶ At 29 U.S.C. § 2102, the WARN Act requires prior worker and governmental notification in the event of a plant closing or mass layoff.¹⁴⁷ 29 U.S.C § 2101(a) defines plant closing and mass layoffs to include, during any 30-day period, a plant closing resulting in employment losses of at least 50 employees; a mass layoff of at least 50 employees where the employment loss consists of at least 33 percent of employment at the site; or a mass layoff with an employment loss of 500 or more at a single site of employment, regardless of its proportion of total employment at the site or if the employment loss is part of a plant closing.¹⁴⁸ Sixty days prior to a termination event that triggers the WARN Act, 29 U.S.C. § 2102(a) requires the employer to give written notice of the planned terminations: “(1) to each representative of the affected employees as of the time of the notice or, if there is no such representative at that time, to each affected employee; and (2) to the State or entity designated by the State to carry out rapid response activities under [the Workforce Innovation and Opportunity Act], and the chief elected official of the unit of local government within which such closing or layoff is to occur.”¹⁴⁹

At 29 U.S.C. § 2107(a), the statute provides the DOL with the authority to “prescribe such regulations as may be necessary to carry out” the act.¹⁵⁰ The DOL recently announced its intention to revise its implementing regulations at 20 C.F.R. Part 639 to update the definition of “single site of employment” as it relates to remote and telecommuting workers.¹⁵¹

The Family and Medical Leave Act (FMLA) at 29 U.S.C. § 2612(a)(1)1 et seq., requires covered employers to offer most employees 12 weeks of unpaid, job-protected leave for the birth and care of a child; to care for an adopted or foster child; to care for a spouse, a child under age 18, or a parent with a serious health condition; or because the employee is unable to work due to a serious health condition.¹⁵²

29 U.S.C. § 2615(a)(1) makes it unlawful for an “employer to interfere with, restrain, or deny the exercise of or the attempt to exercise, any right provided” under the act.¹⁵³ 29 U.S.C. § 265 authorizes the DOL to issue regulations “as are necessary to carry out” the act.¹⁵⁴

Firms already rely on automated systems to manage workforces, a trend that seems likely to accelerate given the proliferation of new AI technologies. But technological innovation does not exempt employers from preexisting statutory obligations. Several statutes empower the DOL to address certain AI issues. GFI and CAP hope this chapter offers inspiration to worker advocates and policymakers interested in how the federal government could update regulatory regimes to account for the ways in which new developments in AI may affect the American workforce.