“I work every other Sunday as a closer, which [ends] at 10:30 or really 11-ish, then [I am] scheduled at 6 a.m. the next morning.” “Managers, if you talk to them about the schedule, they’ll blame [any issues] on the software.” “They expect you to always say yes [to the next shift].”
The quotes from America’s lower-paid employees describe today’s realities for breadwinners whose work hours are decided algorithmically. Quantitative research shows the constant instability. Harvard researchers found “80% of workers have little to no input into their schedules and 69% are required by their employer to keep their schedules ‘open and available’ to work whenever needed.” Many end up with fewer hours than necessary to make ends meet.
Recommended Stories
Analysis of AI’s impact on labor markets often focuses on knowledge workers and whether they will be replaced as they become what one bank executive tactlessly described as “low value human capital.” Those doing physical tasks are regarded as less exposed to AI. Robots are not immediately ready to help an elderly disabled person dress, put up scaffolding, or serve cocktails.
LABOR DEPARTMENT THREATENS TO WITHHOLD FUNDS FROM GOVERNORS OVER UNEMPLOYMENT FRAUD AND WASTE
The impact of AI on this workforce is more about how the technology shapes their hours, pay, and, by extension, quality of life. A corporate employee who might have had a regular-hours job 15 years ago is now likely assigned ad hoc hours with minimal notice through a workforce scheduling system. These platforms run a monopsony labor market for each customer: hundreds, maybe thousands, of staff selling their hours each day to just one buyer, empowered with all the tools and data to achieve business objectives.
This market dynamic heavily favors the buyers — the employers and app developers. It’s good business to minimize labor costs, and AI-based labor markets make that even easier. Consultancies tout the benefits of a just-in-time workforce. McKinsey issued a 2022 report highlighting AI’s capacity to transform “the problem of optimizing schedules so companies can deploy the people they need when they need them and unlock new levels of efficiency.”
Workers may increasingly find themselves locked into this lopsided market. It is tempting to believe workers can vote with their feet, moving to another employer if getting inadequate hours or pay. Or maybe selling their time across multiple gig work apps to maximize income. In reality, it’s hard to be available for work in more than three or four of these systems at once. If you are responding to an assignment from system A when platform B needs you, the algorithms running B will typically downgrade you for unreliability. America’s lower-paid workers are too often left blindly churning through app after app in search of where there might be the most demand and best pay.
There is another way to boost power for blue-collar workers in the AI era: the de-commoditization of labor.
As a supermarket employee in a monopsony labor market for checkout operators, the only way a worker can shine is by being faster at scanning goods or more responsive to their employer’s demands. Likewise, DoorDash workers can only get ahead by delivering more quickly for longer hours. It’s exhausting, demeaning, and a waste of each individual’s multifaceted economic potential.
Swiping groceries and running pizzas are a small part of any human’s unique possibilities for economic activity. An example worker might have a menu of other valuable skills from prior experiences — perhaps their time as a hotel room attendant would qualify them for hospitality roles, or their stint as a babysitter may make them a good fit for childcare-related opportunities.
Could AI unlock possibilities for earning from this wider portfolio of possible contributions to economic growth? Yes. For workers already selling their time hour by hour across multiple sources of employment, AI could maximize earnings and create personalized scheduling. But it requires infrastructure for matching labor demand and supply that can seamlessly connect each person to multiple types of work locally while enabling the collation of data on patterns of demand, supply, and pay as each hour of employment is booked.
There is a sleeping giant with unique leverage and need to deliver that infrastructure, the public workforce system. Often misunderstood by the better-paid, America’s 2,300 job centers are publicly provided alternatives to commercial staffing agencies. Every state workforce agency commissions an all-sectors job-matching platform alongside commercial job boards. It’s all focused on employability, economic growth, data collection, and minimizing public assistance.
But this machine of labor market access, information, and targeted interventions remains solely focused on traditional job creation and placements. It’s as if the 21st century never happened.
But there is a mechanism to allow experimentation that could be fast-tracked as AI impacts breadwinners across the economic base. Labor Department officials can grant waivers to the funding formula, which would allow a state or regional workforce agency to extend its services to also support businesses and residents with more fluid labor arrangements. Novel waivers can be hard to push through the levels of approval. But the department might announce that they would be prioritized to encourage experimentation.
TRUMP NEEDS A PRO-WORKER HEAD OF LABOR DEPARTMENT — NOT A UNION LAPDOG
This wouldn’t be about top-down directives. Simply an announcement that permission to have the outputs of new labor market infrastructure included in annual performance assessments would be fast-tracked. It could spawn a range of possibilities.
As a sign of that, some local boards are already striving to launch their own sophisticated platform for all types of hourly labor across their area, modeled on a U.K. government initiative run by one of the authors. A green light for elected officials and their agencies wanting to offer new options for hourly workers could initiate tools that unleash Americans keen to test their possibilities, explore like-for-like market data, and perhaps shop among competing AI tools dedicated to boosting household income rather than cutting corporate costs.
Professor Kevin Frazier is a senior fellow at the Abundance Institute and the director of the AI Innovation and Law Program at the University of Texas School of Law. Wingham Rowan runs the BeyondJobs.org program, modeled on a U.K. government labor market initiative. Wingham Rowan runs the BeyondJobs.org program, modeled on a UK government labor market initiative.
