Urban Wire How Government Can Embrace AI and Workers
Zach Boren
Display Date

A person's face illuminated by a computer screen.

Artificial intelligence (AI) is set to shake up how we work. But businesses and Congress appear unprepared to meet the moment for workers. While predictions are still nascent, AI will likely have a significant but uneven effect on the workforce—creating new jobs while augmenting others and displacing workers. Studies from OpenAI and the McKinsey Global Institute estimate that AI will affect more than half of job tasks for one-fifth of workers, or automate up to 30 percent of work hours, respectively.

Experts underscored the gravity of this transformation during the Urban Institute’s recent roundtable on AI’s impact on apprenticeships, highlighting a crucial 18-month window for AI adoption and the potentially harmful labor market results if left unchecked. Failing to act during this time could result in harmful outcomes for workers.

Instead of succumbing to panic, policymakers and businesses could use this period could to plan for economic and labor disruptions. Drawing from evidence-based policy during recent economic transformations, we can pave the way for a smoother transition for workers and business.

The state of AI in the workforce

The advancements in productivity ushered in by AI are estimated to increase global gross domestic product by 7 percent. But AI’s introduction to the workplace also creates substantial risks to workers, especially knowledge workers.

AI’s workforce disruption will likely first affect professional jobs, like office support and the legal industry (PDF), where some workers may begin to lose their footing. The disruptions will not likely be limited to a few industries. It’s estimated that 12 million occupational transitions may need to occur in the United States by 2030 because of the evolving nature of work and tech. To harness the potential of an AI-driven future, ensuring worker buy-in requires shared benefits and pathways to good jobs.

How do we achieve AI-inspired economic growth and good jobs? And who bears responsibility for supporting workers and business? Does government or business bear responsibility, or should there be a new tax on AI revenues?

Currently, consensus eludes Congress as a myriad of AI issues—from privacy, intellectual property, worker rights, and disinformation—take center stage. However, there are signs that a congressional strategy regarding AI's effects may be coalescing, with the recent unveiling of the House Bipartisan Task Force on Artificial Intelligence.

Meanwhile, tech leaders are offering and investing in worker solutions. OpenAI’s CEO Sam Altman suggested that the government will primarily pick up the tab for worker transitions. Altman has spearheaded the most significant US investment ($10 million) in universal basic income (UBI), suggesting it as one among several government-led solutions crucial for the anticipated economic upheaval catalyzed by AI. Yet, a new UBI law appears unlikely to gain swift and broad bipartisan support.

Conversely, the White House has committed to advancing AI technology responsibly by issuing an Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. It directs the US Department of Labor (DOL) to “analyze the abilities of agencies to support workers displaced by the adoption of AI,” including current workforce and education programs and possible legislation.

Evolving trade adjustment assistance to AI adjustment assistance

One DOL program offers lessons for how federal policy could adapt to AI while protecting workers.

Over the past few decades, policymakers recognized the trade-offs that liberalized trade agreements created winners and losers in the economy and that some workers would need to reskill and would require income support during the transition. DOL administered a Trade Adjustment Assistance (TAA) program that provided retraining, income support, job search assistance, health care benefits, and even relocation aid.

Since 1974, TAA served more than 5 million workers and enjoyed bipartisan support. Evidence shows that TAA supports a worker transition, with participants earning $50,000 more (PDF) over the 10 years following their displacement. Prepandemic DOL data (PDF) highlight high wage replacement and reemployment outcomes, showing cost-effectiveness. 

However, TAA has also received criticism over the lack of awareness about benefits, challenges in accessing services, narrow eligibility requirements that excluded many workers, and inefficiency and bureaucratic processes (PDF). TAA ultimately lost bipartisan support when it was detached from the Trade Promotion Authority, leaving the worker program unfunded and its future uncertain.

Still, TAA’s successes—and missteps—offer evidence-based lessons for future policymaking. Before Chat GPT, the Senate began thinking of extending TAA to workers hurt by automation with the TAA for Automation Act of 2019. Beyond simply revisiting the bill, Congress could consider implementing the following policies to address job losses and industry changes as a result of AI implementation:

  • Incentives for reskilling: AI response legislation could encourage pre-and registered apprenticeships that align with good jobs. Registered apprenticeships, an industry-designed, earn-while-you-learn training solution, appears to be the strategy most poised to train workers for new AI jobs and skill sets compared with other solutions. Yet, only a few states offer tax incentives for employers offering apprenticeships. South Carolina’s simplified $1,000 per apprentice per year tax incentive (PDF) has helped boost the state’s employers, offering apprenticeships with the potential for national scale.
  • Business and worker supplements: Similar to Short-Time Compensation (PDF) (or work sharing), AI Adjustment Assistance could include partial unemployment compensation to employees facing reduced work hours—averting unnecessary mass layoffs. This approach could help companies retain talent while the federal government replaces a portion of workers’ reduced earnings through unemployment benefits. Policymakers could consider extending unemployment insurance for all AI-displaced workers to allow sufficient time for workers to acquire new certifications and good jobs. Wage insurance (PDF), especially for workers ages 50 and older, could be considered, as demonstrated under TAA (PDF), to support workers where reskilling isn’t a viable option. 
  • Expanded eligibility and oversight: AI Adjustment Assistance would establish eligibility for workers affected by AI, considering job losses caused by increased AI use by employers or significant automation contributing to separations—a similar approach to TAA investigations (PDF). But, to improve eligibility, the federal government could review national or regional labor market data related to occupations or industries adversely affected by AI using AI and other research methods. By considering the scope, individual workers may qualify for benefits, thus responding to previous critiques on worker accessibility.

What an AI Adjustment Assistance Program could look like

When addressing foreseeable labor market disruptions, policymakers can begin by drawing insights into the successes and failures of TAA. But AI and automation present a distinct challenge compared with free trade.

Workers hurt by AI will include low- and high-wage earners. Moreover, AI’s labor market impact on workers won’t discriminate based on political or union affiliation or geography, making this a nascent cross-party crisis. The challenge offers an opportunity to build bipartisan consensus around acknowledging the importance of work, seizing the technological future, and sharing in the economic benefits of greater productivity.

I don’t predict that an AI Adjustment Assistance program will be a panacea to all the AI challenges facing American workers—lawmakers will need to consider a holistic approach to ensure a smooth transition. The National Bureau of Economic Research suggests a broader strategy: that we use AI to shape a better future, tackling the world’s most pressing problems, and that we “blunt or reshape the commercial incentives to use AI for socially counterproductive objectives such as displacing workers” (PDF).

By reconstructing TAA as a life raft for workers affected by AI, federal policymakers can empower businesses to retain and reskill one of our most valuable possessions: their workers and our productivity.


Tune in and subscribe today.

The Urban Institute podcast, Evidence in Action, inspires changemakers to lead with evidence and act with equity. Cohosted by Urban President Sarah Rosen Wartell and Executive Vice President Kimberlyn Leary, every episode features in-depth discussions with experts and leaders on topics ranging from how to advance equity, to designing innovative solutions that achieve community impact, to what it means to practice evidence-based leadership.


Research Areas Workforce Artificial intelligence
Tags Technology, trade, and automation Workforce development Apprenticeships Building America’s Workforce
Policy Centers Center on Labor, Human Services, and Population
Related content