Labor is having a moment in the U.S. The country’s dockworkers just successfully led a three-day strike for higher pay, and crippling strikes by UPS Teamsters and the nation’s rail workers were narrowly avoided in the past two years after concessions from employers and, in one case, intervention from Congress. Public support for unions is at the highest level since 1965, after reaching its nadir in 2009.

The Biden administration also has made a major push to bring key industries like semiconductor manufacturing back to the country—because of both national security concerns and the prospect of creating more good jobs outside of superstar cities. More broadly, the legacy of globalization for American workers is getting a hard look in the presidential campaign and the broader discourse.

Against that backdrop, we’re watching the potential rise of generative AI as the new disrupter of work.

Yes, most credible experts believe change will happen much more slowly and less sweepingly than techno-utopians would have you believe. But also yes, change is already happening.

A major new analysis from the Brookings Institution, using OpenAI data, found that the most vulnerable workers don’t look like the rail and dockworkers who have recaptured the national spotlight. Nor are they the creatives—like Hollywood’s writers and actors—that many wealthier knowledge workers identify with. Rather, they’re predominantly women in the 19M office support and administrative jobs that make up the first rung of the middle class.

“Unfortunately the technology and automation risks facing women have been overlooked for a long time,” says Molly Kinder, a fellow at Brookings Metro and lead author of the new report. “Most of the popular and political attention to issues of automation and work centers on men in blue-collar roles. There is far less awareness about the (greater) risks to women in lower-middle-class roles.”

The Details: The new Brookings analysis stands out among the many analyses predicting AI’s impact because the researchers had access to OpenAI data. Kinder says this allowed them to do a much more fine-grained analysis of the specific job tasks that can already be done by generative AI. That in turn allowed the researchers to analyze not just exposure to AI—which could mean anything from augmenting a job to eliminating it—but the specific risk of having your role fully automated. (In an interesting turn, the human-led research for the Brookings report was augmented by data analysis performed by ChatGPT.)

The researchers found that:

  • More than 30% of all workers could see at least 50% of their occupation’s tasks disrupted by gen AI.
  • Higher-paying jobs in fields like computing, business and finance, engineering, and law are highly exposed to the impacts of gen AI, but the technology is most likely to augment or take over discrete tasks rather than entire jobs.
  • On the other hand, lower-paying office jobs like insurance claims processors, legal secretaries, bookkeepers, and administrative assistants are now much more likely to see their roles fully automated.

It may already be happening for some job roles. Salesforce last month rolled out new AI-powered agents, dubbed Agentforce, that are capable of working autonomously. The company is billing the service as a way for customers to augment their existing workforces but also says that could mean hiring fewer seasonal employees or gig workers during busy times. 

To help workers keep up, Salesforce is investing $50M in training facilities and making AI courses offered through its Trailhead platform free.

“We firmly believe that AI represents the biggest platform shift we’ve ever seen,” John Somorjai, chief corporate development and investments officer at Salesforce, said on a recent call with reporters. 

Many administrative jobs have already been eaten away by software-based automation and earlier versions of AI—but these roles still employ a large chunk of the labor force. And they represent the lion’s share of “decent-paying, stable jobs” for women without a college degree, the Brookings researchers write.

A Seat at the Table: Kinder and her co-authors go to great lengths to make clear that they can’t predict what’s going to happen. But their report argues that government and industry leaders need to have much more serious conversations about what’s already changing and what they envision going forward—and that the workers whose jobs are most likely to change need to be a part of that conversation. 

U.S. Secretary of Commerce Gina Raimondo has floated going much further. She’s proposed a kind of grand bargain between employers and workers that would provide employees with some sort of industry-wide job guarantee in exchange for an openness to experiment with AI and to retrain when necessary. 

“They have a sense of security, which they deserve,” Raimondo said on the Possible podcast with Reid Hoffman. “They have a job, which they deserve. And then everyone else is a bit more open to pushing forward with the possibilities of AI.”

Thus far, it’s just an idea. But Raimondo is perhaps the most prominent among a rising chorus of voices on both the left and right who are calling for workers to have a bigger seat at the table—one of several key trends we outlined in a recent Work Shift explainer on “AI in Its Play Era.”

Mary Alice McCarthy, senior director of the Center on Education and Labor at New America, says that bringing industry and labor together isn’t just about protecting livelihoods. “A lot of automation is so-so—and workers can help make it better,” she says.

That’s at least part of the theory behind burgeoning partnerships like the one between Microsoft and the AFL-CIO, the nation’s largest federation of labor unions. But the vast majority of workers in the U.S., and especially those in office jobs, aren’t represented by unions. And, Raimondo’s comments aside, state and federal policymakers have been slow to bring industry-worker conversations to the fore.

This creates what Kinder and her colleagues dubbed the “Great Mismatch,” meaning that the workers who most need a seat at the table are the least likely to have a clear mechanism to ask for it. “The vast majority of workers in sectors that will likely experience the most disruption from AI are not represented by formal unions or other forms of worker voice,” Kinder says.

The task now, she says, is for policymakers and other leaders to figure out how to build those mechanisms. And the next two years will be critical.