Understanding the DOL’s Guidance on Using AI for FMLA Management

AbsenceSoft
September 26, 2024

HR practitioners are deeply familiar with the work it takes to manage leave under the Family and Medical Leave Act (FMLA). From fielding requests and calculating eligibility to collecting documentation and facilitating return to work, the FMLA leaves HR with an abundance of tasks.

Many marketplace solutions are using artificial intelligence (AI) to reduce HR’s workload — everything from streamlining recruitment and hiring to recommending benefits for employees. Providers tout the benefits of AI and reason that the technology can execute straightforward, administrative tasks — the stuff FMLA management is made of, they say.

It is true that much of the FMLA process can be automated, just as our platform does, and the time savings are always significant. However, for some parts of the FMLA, and AI technology don’t make a great match. The FMLA can be a complex statute that requires careful review, or you run the risk of a complaint or a costly lawsuit.

As for AI, its limitations can make it a risky tool when it comes to something as nuanced as making decisions and calculating entitlements in leave management. What’s more, it can be challenging to ascertain how HR vendors are using AI if you don’t know the right questions to ask.

We’re not the only ones highlighting potential limitations on AI in FMLA management. The U.S. Department of Labor (DOL), the agency that enforces the FMLA, recently published guidance explaining why the use of AI in FMLA management creates the potential for widespread, systemic compliance violations.

Below, we’ll dig into the complexities of FMLA management and explore DOL’s guidance. We’ll conclude with a few pointers on how to evaluate AI vendors and tools to minimize risk and maximize productivity in your organization.

The DOL opened its guidance by acknowledging AI’s capabilities. The technology can streamline tasks, improve efficiency, and strengthen workplace safety, the agency pointed out. The DOL’s point is an important one to make: AI is a powerful tool.

What does the DOL say about using AI to manage FMLA?

After highlighting the benefits of AI, the DOL went on to stress the importance of human oversight, especially where AI is involved with federal labor law. Without human involvement, AI has the potential to create organizational-wide compliance challenges, the DOL said.

“As new technological innovations emerge, the federal laws administered and enforced by [the DOL’s Wage and Hour Division] continue to apply, and employees are entitled to the protections these laws provide, regardless of the tools and systems used in their workplaces.”

The agency emphasized that such compliance issues don’t just expose organizations to legal risk. They could threaten workers’ rights.

The DOL showed how, for instance, an AI timekeeping program can incorrectly interpret an employee’s inactivity at a computer as time off work. The employee may have been performing work away from the computer — in a meeting or at a coworker’s desk, perhaps. If the program docks that employee, it’s stripping them of the wages they rightfully earned under the Fair Labor Standards Act.

The unsupervised use of AI carries similar risk when it comes to FMLA management. Employers may deny workers the leave they’re entitled to if an AI system errs in calculating their eligibility. This scenario isn’t a distant possibility.

To correctly determine an employee’s entitlement to FMLA and other leaves, HR must carry out detailed administrative work and make complex judgment calls. It’s a job that necessitates human involvement. It’s not something that can be entrusted to AI alone.

Consider the case of Wendy. Wendy is a healthcare worker in California who has been at her job for five years. She recently delivered a healthy baby. Her leave case may appear simple on the surface, but the details of her pregnancy reveal vast complexities.

Wendy struggled to get pregnant. When she finally conceived, her pregnancy gave way to complications that required bedrest before delivery. Once the baby arrived, Wendy experienced a few complications as she recovered, including postpartum depression.

Wendy’s case is anything but simple. She is entitled to FMLA leave, of course, but that’s one of many laws granting her protections. In total, Wendy has access to:

Federal Laws

State Laws

  • The California Family Rights Act
  • California Pregnancy Disability Leave
  • California’s Paid Family Leave (PFL)
  • California State Disability Insurance

Company Policies and Benefits

  • PTO (paid time off)
  • Paid sick time
  • Parental leave programs
  • Short-term disability
  • Union programs

A case like this one highlights the vast, overlapping complexities of FMLA management. It’s easy to see how AI could struggle to navigate the complex interactions between all of Wendy’s entitlements. And an incorrect decision in several of those entitlements, including FMLA, could lead to a complaint or even a lawsuit. It could also make Wendy’s leave experience more stressful by repeatedly asking for too much documentation or miscalculating her return-to-work date.

AI can be extremely helpful in making FMLA management a more streamlined process. But it cannot be used without human oversight, as the DOL emphasized. To take advantage of all that AI has to offer without breaching the DOL’s guidance, follow the tips below.

Tip #1: Don’t delegate entitlement and eligibility calculations to AI

When employers rely on AI to process leave requests, track time off, and integrate absence calendars, they may encounter serious compliance issues. Timekeeping programs can incorrectly determine hours worked, which may cause errors in eligibility calculations.

The DOL also noted that an automated system testing for eligibility may do so more frequently than permitted under the FMLA, which could also result in employees’ leave being denied incorrectly.

The DOL acknowledged that human leave managers can make these mistakes as well, but they are often a single mistake–not a widespread, repeated, systematic error. When AI tools and algorithms make decision errors, they can easily “could result in violations across the entire workforce,” according to the DOL.

Tip #2: Don’t let AI handle medical certifications on its own

Under the FMLA, employers are allowed to collect only as much medical information as is necessary to process the leave. AI systems that handle medical certifications may prompt excessive or unnecessary requests for employee information, the DOL said. Employers should make sure systems “flexibly accommodate certification deadlines without penalizing employees,” the DOL said.

This flexibility is important when it comes to processing medical certifications under varying leave laws. Remember, AI looks for patterns. If it applied FMLA “rules” regarding medical certification to a law like the Pregnant Workers Fairness Act, employers could land in hot water. The PWFA differs from the FMLA on medical certification; it holds that a conversation with an employee will provide sufficient information about a worker’s limitation, making documentation unnecessary.

Nuances like these require human oversight in leave management. Otherwise, employers may find themselves facing the very problems the DOL warns about in its guidance — systemic compliance problems carrying “significant legal challenges.”

Tip #3: Watch for interference, retaliation, and discrimination

The DOL noted that some employers use AI systems to track and analyze leave use. Employers may not use such systems to retaliate against workers who take leave or discourage employees from taking leave.

The DOL went on to say that employers may not use AI to count FMLA leave as a negative factor in employment decisions. Employers cannot, for instance, view FMLA use as a reason to deny someone a job or a promotion. FMLA use also may not be a reason to take disciplinary action against an employee.

The agency also wrote that an employer would be in violation of the FMLA if an AI tool failed to provide benefits to a worker on FMLA leave (assuming the employer provides those benefits to workers on similar types of leave).

Again, the DOL warned employers about the widespread nature of potential errors caused by AI. Unlawful retaliation or interference, it argued, could “occur across a large group of employees.”

Implementing AI the right way

The DOL is right to highlight the errors an AI-powered system may create in FMLA management. However, those errors only arise when AI is used without human oversight. When implemented responsibly, AI and other advanced technologies can streamline HR’s workload without increasing the risk of compliance mistakes.

If you’re considering adding AI to your leave process — FMLA or otherwise — you’re probably searching for software providers. Begin your search by deciding where you want to implement AI. Evaluating a vendor’s use of AI with a greater understanding of AI’s benefits and drawbacks will help you determine the right tool for the job and compare various solutions and providers.

This step is critical to creating or maintaining a leave management program that’s compliant and compassionate. Repeatable, straightforward, administrative tasks are well suited for automation. By automating these tasks, HR saves time so they can provide personalized, one-on-one support to employees.

Conversely, calculations regarding employee eligibility and entitlements should hinge on human review and oversight, as AI may struggle with complex leave scenarios. Most importantly, key decisions carrying significant compliance implications should never be made by AI.

Technology should not, for instance, approve or deny someone’s leave request. These decisions should always be made with human input. Once you’ve decided where you’d like to incorporate AI, begin to evaluate vendors and the technology they offer. Consider the following checklist:

  • Solutions should NOT allow AI to make key decisions.
  • Technology offerings should be fully auditable.
  • The vendor should actively review for systemic bias.

A leave platform that embraces AI — but not its risk

The AbsenceSoft platform incorporates automation that streamlines HR’s workflow, minimizes administrative labor, and improves the employee experience If you’re on the lookout for a leave solution that harnesses the power of automation with a mindful, compliant approach, schedule a demo with AbsenceSoft.