In today’s discourse surrounding workplace policies, the term “algorithmic management” is omnipresent. We are frequently warned that companies are wielding an assortment of digital “tricks” to exert control over their employees. They monitor our every keystroke, track our movements via GPS, and even peer through our webcams. This narrative has permeated academic journals, government reports, and mainstream media. It has even found its way into federal regulations, particularly within the U.S. Department of Labor’s latest rules on independent contractors. Much like the sensational tales that dominate the headlines, this rule presumes the widespread presence of algorithmic management, framing it as a form of “control.”
However, there’s a catch: algorithmic management, as it stands, is a misnomer. This term is employed to describe age-old business practices that would otherwise be considered mundane if not draped in an ominous label. These practices encompass performance monitoring, incentive provision, and work tracking. Critics have deftly rebranded these established methods as novel and detrimental to the American workforce, and this misleading narrative has been influencing policy at the Department of Labor for several years.
Fortunately, this narrative may be losing steam. The Department of Labor has recently proposed to remove algorithmic management from its regulatory language, opting instead for a more conventional understanding of workplace “control.” This redefinition would implicitly acknowledge that monitoring and incentives are not inherently forms of control; rather, they represent how businesses manage external suppliers when direct oversight is absent. In this light, these practices do not signify control, but rather the absence of it—even when they are cloaked under the foreboding title of “algorithmic management.”
Algorithmic Management: An Empty Concept
The term algorithmic management has always lacked clarity. It was first introduced in a 2015 academic paper centered on the then-emerging rideshare platforms. The study explored how these platforms employed digital tools to coordinate a vast network of independent drivers, focusing on three primary tools: matching riders with drivers, customer ratings, and surge pricing. The conclusion was that these tools not only helped balance driver supply with rider demand but also contributed to driver satisfaction. Surprisingly, algorithmic management appeared to be a positive force.
Since then, however, the term has morphed into a catch-all for various tools, from scheduling software to AI-driven chatbots, with the latter inciting heightened fears. Writers such as Noam Scheiber of the New York Times, Sam Levine from the Federal Trade Commission, and Veena Dubal of UC Irvine have painted algorithmic management in a sinister light, suggesting that companies use it to manipulate workers into longer hours for lower pay. Dubal has even gone so far as to liken it to a modern Jim Crow.
These claims are not merely rhetorical flourishes; they have begun to shape public policy. In 2024, the Department of Labor implemented a regulation that delineates the distinction between employees and independent contractors, a critical differentiation as only employees are entitled to minimum wage and overtime protections under the Fair Labor Standards Act (FLSA). The 2024 regulation expanded the definition of control to encompass certain algorithmic management techniques, specifically citing “technological” monitoring, such as GPS tracking, as a form of control, even in isolation.
This shift marked a departure from the department’s traditional interpretation of control, which required affirmative actions from the business—namely, direct instructions, prohibitions, or punitive measures against workers. The new rule broadened this concept to include passive observation, suggesting that simply collecting work-related data via technology could be construed as controlling that work.
Analog Reasoning in a Digital Age
This line of thinking misrepresents the issue at hand. Incentives, monitoring, and related techniques do not signify control; they indicate its absence. Economically speaking, these are mechanisms businesses employ to address the “principal-agent problem.” The crux of the issue lies in the divergent incentives between a business and its contractors: while the business desires the best service at the lowest cost, the contractor aims for the highest compensation for the least effort. Consequently, businesses must adopt measures to safeguard against subpar performance. If they were hiring employees, they could manage this risk through direct instruction. However, the nature of contracting often necessitates a more indirect approach, such as milestone reporting (monitoring) or performance bonuses (incentives).
This indirect influence has historically never been deemed “control” for worker classification—and it shouldn’t be. If it were, distinguishing any work as independent would be an uphill battle. After all, all principals monitor their agents to some extent; therefore, monitoring alone provides no clarity regarding the classification of a relationship. For instance, when a company ships a package and requests delivery confirmation, it is, in a sense, monitoring the work, but no one would argue that all logistics suppliers should be classified as employees of their clients.
Modern technology does not alter this fundamental dynamic. Businesses and contractors still grapple with principal-agent problems, and the most effective methods for navigating these challenges remain monitoring, incentives, and similar indirect strategies. While these techniques may function more swiftly in the digital realm, the underlying dynamics remain unchanged—even when they are branded as “algorithmic management.”
Indeed, the very efficiency of these algorithmic tools is what fuels public apprehension. Many fear that if an employer can monitor keystrokes or access webcams, they might exploit this information to micromanage employee performance. Yet the crux of the matter is not whether information is collected, but how it is utilized. If an employer leverages keystroke data for evaluation, compensation, or disciplinary measures, it is the application of that data, not its collection, that constitutes control. Legally speaking, the focus should be on whether the principal exerts control over the work—not whether they are aware of how the work is conducted. This principle holds true regardless of whether the information is gathered through technology, a delivery confirmation, or direct observation.
Luckily, the Department of Labor appears to be aligning with this perspective. In late February 2026, it introduced a new rule that clarifies the distinction between contractors and employees, omitting any reference to algorithmic management or other forms of “technological” control. Instead, it reverts to fundamental principles: if an individual controls their own work, they are likely a contractor; if they do not, they are likely an employee. This conclusion remains consistent even as the nature of work evolves in the digital landscape.
