AI in HR: DMEC Urges Guardrails for Workplace Leave Decisions

SAN DIEGO, CA – February 25, 2026 – As artificial intelligence quietly reshapes the modern workplace, a national nonprofit is sounding the alarm on its use in one of the most sensitive areas of employment: managing worker leave and disability accommodations. The Disability Management Employer Coalition (DMEC) today released a first-of-its-kind policy brief urging state policymakers to establish clear rules for AI systems that are increasingly making decisions affecting employees' health, income, and job security.

The brief, Artificial Intelligence in Workplace Leave and Accommodation Management, arrives as employers rapidly adopt AI to streamline complex HR functions. These tools promise efficiency in determining benefits eligibility, processing claims, and monitoring compliance. However, DMEC warns that without robust guardrails, these same systems risk introducing new forms of bias and discrimination, creating opaque decision-making processes that could harm vulnerable workers.

"When AI is used in systems that directly affect a person's health, income, and job stability, early and thoughtful policymaking matters," said Bryon Bass, CEO of DMEC, in a statement accompanying the release. Bass stressed the goal is to equip policymakers with the information needed to "support innovation while ensuring fair and responsible workplace practices."

The Double-Edged Sword of AI in HR

For human resources departments buried in paperwork and complex regulations, AI offers a compelling solution. Vendors now offer platforms that can automate eligibility checks for federal and state leave laws, generate required notices, and track intermittent leave with precision. Proponents claim these systems can reduce administrative time by up to 40%, minimize costly compliance errors, and free up HR professionals to focus on more strategic, human-centric tasks.

But this efficiency comes with significant risks. The DMEC brief highlights that AI tools are only as good as the data they are trained on and the logic they follow. If an algorithm is trained on historical data that contains implicit biases, it can learn to perpetuate and even amplify discriminatory patterns. This could lead to an AI system unfairly denying a leave request or failing to suggest a reasonable accommodation for a qualified employee with a disability.

The challenge lies in the technology's complexity. Many AI systems operate as "black boxes," making it difficult for employers to understand, let alone explain, how a specific decision was reached. This lack of transparency poses a major hurdle for accountability and can leave employees with little recourse to challenge an adverse, algorithm-driven outcome.

A Rising Tide of Regulation

DMEC's call for action lands in an already active but fragmented regulatory environment. Lawmakers and agencies across the country are grappling with how to govern AI in employment, resulting in a patchwork of rules that creates a compliance minefield for national employers.

At the federal level, the Equal Employment Opportunity Commission (EEOC) has repeatedly affirmed that long-standing civil rights laws, like the Americans with Disabilities Act (ADA) and Title VII, apply fully to AI-driven employment decisions. The agency has made AI bias a top enforcement priority, placing the legal responsibility squarely on employers to ensure their automated tools are not discriminatory, regardless of whether they were developed in-house or purchased from a third-party vendor.

State and local governments have been more prescriptive. New York City's Local Law 144, which took effect in 2023, requires employers using AI for hiring and promotion to conduct annual independent bias audits and notify candidates about the technology's use. In a more comprehensive move, Colorado's Artificial Intelligence Act, set to take effect in February 2026, imposes duties on both developers and users of "high-risk" AI systems to manage risks of algorithmic discrimination.

California has also moved aggressively, with new regulations effective in late 2025 that hold employers liable for discriminatory impacts from third-party AI tools and require detailed record-keeping. This growing web of disparate state laws underscores the need for the kind of cohesive policy guidance DMEC is proposing, which aims to create a more standardized framework for responsible AI use.

When Algorithms Discriminate

The theoretical risks of AI bias have already materialized in real-world applications, leading to high-profile failures and legal challenges. Amazon famously scrapped an AI recruiting tool after discovering it systematically penalized resumes containing the word "women's," as it had been trained on a decade's worth of predominantly male resumes. More recently, the EEOC settled a case with a tutoring company for $356,000 after its hiring software was found to have automatically rejected older applicants.

These examples from the hiring context provide a stark warning for the leave and accommodation space. An AI tool designed to assess an employee's fitness for duty could, for instance, penalize an applicant for a gap in their employment history, inadvertently discriminating against someone who took legally protected medical leave. Similarly, an algorithm analyzing performance metrics might fail to account for a reasonable accommodation provided to an employee with a disability, leading to an unfair and illegal evaluation.

Legal experts caution that the use of flawed AI is becoming a significant source of employment litigation. A closely watched class-action lawsuit against software provider Workday alleges its AI-powered screening tools discriminate against applicants based on age, race, and disability, highlighting the immense liability employers face when deploying these systems.

A Call for Human-Centered Guardrails

Rather than advocating for a ban on AI, the DMEC brief provides a forward-looking roadmap for policymakers to foster responsible innovation. Central to its recommendations is the principle of human oversight. The brief calls for requirements ensuring that a human being reviews any high-risk or consequential decision made by an AI, such as the denial of a leave request or an accommodation.

Other key policy considerations outlined include establishing clear governance structures for AI systems, mandating transparency through disclosure of AI use to employees, and implementing strong data security standards to protect sensitive health information. The brief encourages the development of ethical AI guidelines and robust protections against algorithmic bias, pushing for systems that are not only efficient but also equitable and transparent.

By releasing this framework, DMEC aims to shift the conversation from a reactive posture to one of proactive planning. The organization is making the brief and other resources available to legislative officials through its State Policy Resource Hub, inviting collaboration to build a future where technology supports, rather than subverts, fair and compassionate workplace practices.