AI and the shifting landscape of exempt employee classification

Artificial intelligence is no longer a novelty in the workplace -- it is rapidly becoming embedded in how employees perform their jobs and how employers define expectations. Across industries, AI-enabled tools are reshaping workflows, timelines and even the skill sets required for success.

Public debate often frames AI as a bogeyman poised to replace workers altogether. Setting that important debate aside, there can be no doubt that its use across industries is changing the work that employees do. Yet these changes can run headfirst into federal and state wage and hour laws. When it does, it opens the door to claims that the exemptions do not apply, that the exempt-classified employees have been misclassified, and that their employers owe them unpaid overtime and, in California, meal and rest period premiums.

FLSA and California law exemptions

The white-collar exemptions -- executive, administrative and professional -- require, in addition to their compensation requirements, satisfaction of a qualitative primary duty test that turns, as the name suggests, on the character of the duties that an employee performs. California's white-collar exemptions focus on the duties that an employee performs, too, but also test the quantitative basis -- if an employee devotes more than 50% of their work to nonexempt duties, then they cannot be exempt.

AI eliminating and augmenting duties

When AI begins to eliminate or reduce managers' direct responsibility to make choices of who to interview for open roles, who to hire, how to evaluate subordinates, and when to reduce workforces, those managers' exempt status might be placed at risk. If an AI tool evaluates data, weighs options, and generates a recommended course, the employee's role may shift from decision-maker to implementer. The more that human judgment is replaced rather than informed by AI, the weaker the exemption argument becomes.

Similarly, AI platforms are already adept at creating videos, drawing pictures, telling jokes, and writing articles -- stepping into the work of creative professionals across professions like marketing, design and journalism. If a fashion designer relies exclusively on AI to design next year's fashion line, or if a journalist uses AI to prepare this Sunday's report on the next election cycles' congressional candidates, how might that risk those employees' exempt classification?

And among AI's most impressive feats is its ability to develop complex computer coding, making programmers lives a little easier. But if that timesaving functionality eliminates their primary role as coders, it could also be placing their exempt status at risk.

Evident from all of this is that when the use of AI replaces some component of an employee's job, it increases the risk that it will also eliminate the justification to classify the employee as exempt. This is especially true in California, where efficiencies reducing the quantity of time spent on exempt duties might affect exempt status no matter the quality of the work being done.

Litigation and enforcement trends

Remarkably, there are no reported decisions -- yet -- determining how AI might affect employees' exempt classification. The DOL's Wage and Hour Division has discussed how AI might affect workers' jobs and flagged compliance risks around scheduling, timekeeping and monitoring. But no guidance has been issued on the question of AI's impact on questions of exempt classification.

One pre-AI-revolution case -- Lola v. Skadden et al., 620 F. App'x 37 (2d Cir. 2015) -- might provide a hint of the issues with which employers and courts will grapple as AI continues to proliferate in the workplace.

In Lola, a former contract lawyer filed an FLSA collective action against Skadden, claiming that his work on document review projects was work that a machine could have done. After the district court granted Skadden's motion to dismiss his complaint, finding that his status as a lawyer exempted him from overtime pay protections under the learned professional exemption, the 2nd Circuit reversed, concluding that "an individual who . . . undertakes tasks that could otherwise be performed entirely by a machine cannot be said to engage in the practice of law" and -- at least at the pleadings stage -- it was conceivable that work as a document reviewer might create just that predicament.

No doubt more cases and further agency guidance are on the way.

Risk mitigating measures for employers

It is also no doubt that there are new wage and hour claims that will be brought in this new AI era.

For employers and their advisors, there are steps that might be taken to mitigate the risk that those claims will bring with them.

Employers should:

1. Map duties to exemption tests to see where discretion currently exists.

2. State in policies that AI outputs are advisory only.

3. Train employees to evaluate AI recommendations critically.

4. Configure AI systems to require human approval.

5. Keep evidence of human involvement in key decisions, i.e., keep "humans-in-the-loop."

6. Audit regularly to ensure duties match exemption criteria.

Conclusion

AI is changing employees' work and employers' expectations for how work is done. It's a new world in which AI can positively impact employee efficiencies, efficacy and accuracy. Employers should be cautious, however, to monitor whether those positive changes run the risk of eliminating the need for exempt-classified employees to do the work justifying their exempt classification. The lawyers who practice wage-hour law, and the courts before whom they practice, should prepare for the likely waves of litigation sparked by employees' use of AI.