Can the Fight Against AI Revitalize the US Labor Movement?

Not long ago, the National Eating Disorders Association (NEDA) summarily dismissed the staff of its helpline after members tried to organize a union. Managers at the nonprofit then tried to replace the helpline with an AI-powered chatbot, according to an account by Abbie Harper for NPQ, one of the leaders of the unionization effort.

Recent developments suggest that the labor movement is starting to take a more active and adversarial approach toward AI in the workplace.

This was hardly an isolated incident amid the breakneck integration of artificial intelligence (AI) into the workplace. Ominous signs abound, from the news editor at a nonprofit news site who used AI to edit stories and then fired a reporter for raising objections, to Salesforce’s cutting 4,000 customer support jobs and shifting to AI agents.

However, the incident at NEDA captures two stark realities of the current economic and technological moment: an enthusiastic and essentially unregulated embrace of AI on the part of employers on the one hand; and a weak labor movement on the other. Will this dynamic bring about a labor resurgence? It was, after all, an earlier fight for worker dignity amid rapid technological change—the Industrial Revolution—that brought the union movement into existence more than a century ago.

Recent developments suggest that the labor movement is starting to take a more active and adversarial approach toward AI in the workplace, going beyond the Hollywood strike in 2023, when unions representing writers and actors won concessions from major studios limiting how AI can be used to augment or supplant their work.

“The topic of AI and other digital technologies are showing up in contract fights and organizing campaigns, increasingly in public policy debates across a variety of sectors,” Lisa Kresge, senior researcher at UC Berkeley’s Labor Center, told NPQ. “It really has in the last few years been at an inflection point, which reflects where we are in the moment of AI, as well as where unions are in recognizing the potential threat of AI.”

Some union leaders and organizers seem to view the fight against the excesses of AI as a way to vault the labor movement back into relevance.

Wading into the Fight

Last October, the AFL-CIO issued its first set of principles aimed at protecting workers in the age of AI.

“There is an urgency to have a meaningful national conversation about how to both propel innovation and adopt sensible policies that protect working people and the general public from the well-documented negative consequences of unregulated AI,” the document states.

“Across all unions the foundational concern is that they’re given no notice or transparency or disclosure about the employer’s plans to adopt the technology.”

The AFL-CIO lays out eight key principles for protecting workers, including “guardrails against harmful uses of AI”; support for copyright and intellectual property protections; a “worker-centered workforce development and training system”; and “transparency and accountability in AI applications.”

Other major US-based unions have weighed in on the issue in recent months.

“Every day, millions of American workers play by rules they never wrote, rules that all too often fail to consider the human aspect of work,” wrote April Verrett, president of the Service Employees International Union (SEIU), in a post for the Aspen Institute.

Verrett cited examples like home healthcare workers “whose algorithm-assigned schedule leaves no time between clients” and security guards “whose digital tracker penalizes bathroom breaks.”

AI at the Bargaining Table

Union engagement on these issues goes beyond rhetoric. AI has become a key point in negotiations between unions and employers across a range of sectors. A search of a database kept by the UC Berkeley Labor Center shows nearly 100 contract provisions addressing AI, including between the Associated Press (AP) and the Communications Workers of America (CWA); Boeing and the AFL-CIO; and General Motors and the United Auto Workers.

The provisions of the agreements vary by sector, but they share some broad priorities. One of them is the principle that humans must be closely involved in the use of AI technologies, sometimes referred to as “human in the loop.”

The AP-CWA agreement, for example, states: “Generative AI may be used to perform the work of news production only with the direct involvement and oversight of employees in compliance with AP standards.”

Another common provision in union agreements seeks to place limits on algorithmic management—that is, using automated technologies to monitor, evaluate, and even discipline employees.

Perhaps the most consistent AI priority across industries is the requirement that employees be notified when employers plan to bring such technologies into the workplace.

“Across all unions the foundational concern is that they’re given no notice or transparency or disclosure about the employer’s plans to adopt the technology,” Kresge told NPQ. “They’re basically blindsided by new systems that are introduced into their workplaces.”

A Nuanced Approach

For all the heated rhetoric around AI, unions’ approach to the technology can hardly be described as militant or even entirely adversarial. Some unions, including the AFL-CIO, have collaborated with tech companies and AI researchers to devise AI tools with worker protections in mind. And it bears noting that many nonprofits are adopting AI within the framework of ethical AI policies, with the goals of enhancing their missions and reducing the workload of often overburdened staff.

UNITE HERE, which represents about 300,000 workers in the hospitality sector, has worked closely with researchers at Carnegie Mellon University to develop AI systems that incorporate the perspectives and priorities of frontline customer-facing staff.

“The thing that we have observed is that, while there is some job replacement [with AI], there is a lot more job transformation, at least in hospitality,” Ben Begleiter, a longtime UNITE HERE leader who has been closely involved in the union’s work on technology, told NPQ.

Under agreements negotiated by UNITE HERE, employers must, among other things, notify and negotiate with union representatives before implementing AI technologies that affect them. Employers must also provide support to workers whose jobs are lost due to AI.

“There’s a recognition that we’re never going to save every job, and so there’s severance pay and extended layoff language…to help cushion the blow or provide for additional opportunities,” Begleiter said.

Beyond Organized Labor

Unions that have increased engagement around AI run up against a sober reality, that they are merely shadows of their former selves, representing just under 10 percent of US workers—half the share they represented in the early 1980s.

The weakening of the union movement coincides with a trend that in many ways is its antithesis: the growth of the gig economy, which has been driven by technology platforms like Uber and Amazon. Not only is the sector largely impervious to unionization, but it also remains an unsettled question whether gig workers are even protected by existing employment laws.

With algorithmic hiring and monitoring baked into the platforms from their inception, gig workers may be canaries in the coal mine when it comes to how AI could influence future work practices.

What recourse do gig workers—and the vast majority of workers not represented by unions—have in the face of undisclosed and abusive use of AI in the workplace?

“We are seeing a coalition of folks that are starting to come together to make some real clear demands around how these tools are used in and on our communities.”

Working Partnerships USA is among several organizations around the country seeking to give workers greater power outside of the traditional union framework. Based in San Jose, CA, in many ways in the belly of the beast of the “new economy,” they’ve organized several campaigns, including Gig Workers Rising and Silicon Valley Rising, aimed at unifying workers and giving them more clout or at least a voice at the table.

Maria Noel Fernandez, Working Partnerships USA’s executive director, told NPQ that she senses a kind of quiet groundswell around issues related to AI:

I think particularly under the current administration, we are seeing gig workers unite with the immigrant rights groups, unite with folks that are worried about the impact on our criminal justice system, unite with folks afraid of what AI is doing to make decisions around health, to parents and teachers worried about what AI is doing in our classrooms.

Fernandez added, “While the corporations are big and have certainly fought back in California, I think we are seeing a coalition of folks that are starting to come together to make some real clear demands around how these tools are used in and on our communities.”

Preserving jobs and worker dignity amid the latest wave of AI technologies sweeping workplaces will require new forms of organizing and collective action. But traditional unions could still play a vital role, one that goes well beyond their diminished ranks. After all, the victories won by unions on work conditions and pay tend to have a “spillover” effect.

At the very least, Kresge of UC Berkely says, the mobilization around AI and worker rights could be a catalyst for a long-overdue public debate on how AI can be used to make jobs more fulfilling rather than more precarious.

“There’s a lot of research that indicates that the rights and standards that unions have negotiated in collective bargaining,” Kresge said, “have eventually either become law for workers across the country in various cases or are setting basically the standards by which employers are competing.”