The Dangerous Drift in Democratic Thinking on AI and Jobs

Good ideas on AI and work are taking shape, but not every proposal will deliver

Artificial intelligence (AI) is moving faster than almost any other issue, and the pressure on lawmakers to respond is mounting. The good news is that Democrats aren’t starting from scratch. Across both chambers, members have introduced legislation that points toward a serious, worker-first approach to AI, focused on training, education, research infrastructure, and better data on how AI is actually affecting jobs.

But alongside that real progress, some ideas gaining traction on the left would either fail to deliver for workers or actively undermine the gains AI could produce. Getting the future of work right means doubling down on what’s already working and steering clear of proposals that miss the mark.

Where Democrats are making strides

1. Job-connected training

The bipartisan AI Workforce Training Act, led by Representative Josh Gottheimer (D-NJ-5), recognizes a simple yet important reality: most training decisions happen inside firms rather than in government programs. By offering tax credits to employers that invest in AI-related career development, including accredited courses, workshops, in-house training, and practical skills like data literacy and prompt engineering, it pushes companies to upskill their own workers rather than replace them.

Sen. Mark Kelly’s (D-AZ) AI for America roadmap complements this with calls for apprenticeships, community college credentialing, and employer-provided training, along with expanded individual training accounts that put upskilling resources directly into workers’ hands.

2. AI education

Bipartisan efforts like the NSF AI Education Act, introduced by Senator Maria Cantwell (D-WA) and Representative Andrea Salinas (D-OR-06), seek to expand access to AI training and scholarships, particularly for those traditionally excluded from the tech industry.

Senators Lisa Blunt Rochester (D-DE), Mazie Hirono (D-HI), and Adam Schiff (D-CA) have also introduced legislation, the Workforce of the Future Act, which would authorize up to $160 million in Education Department grants to expand access to AI education, including teacher training and recruitment.

Recent proposals from Rep. Suzanne Bonamici (D-OR-01) emphasize a “human-centered” approach to AI education, including investments in teacher training to equip educators to bring these concepts into the classroom.

A bipartisan bill led by Senator Adam Schiff (D-CA), the Literacy in Future Technologies Artificial Intelligence (LIFT AI) Act, further builds on these efforts by focusing specifically on K-12 AI literacy. This legislation would support National Science Foundation (NSF) grants to develop AI curricula, expand professional development for educators, and promote hands-on learning to help students understand and use AI technologies responsibly.

3. AI research infrastructure

Recent bipartisan efforts from Senators Maria Cantwell (D-WA) and John Hickenlooper (D-CO) propose expanded federal investment in AI research infrastructure, including testbeds, standards development, and public-private partnerships to accelerate deployment. This is a foundation that future job-quality initiatives can build on.

4. Measuring AI’s impact on jobs

Sen. Mark Warner (D-VA) and Sen. Josh Hawley (R-MO) have pushed to strengthen the way the Department of Labor tracks AI’s impact on the workforce, including improvements to core datasets such as the Current Population Survey and the Job Openings and Labor Turnover Survey. Their proposed AI-Related Job Impacts Clarity Act (S.3108) reflects a broader effort to expand reporting on hiring, job loss, and direct displacement due to AI to better understand its labor market effects.

Taken together, these efforts show that Democrats are already beginning to build the infrastructure for a worker-first AI agenda. The task now is to expand on them – and avoid the traps that could hinder their impact.

Avoid policies that miss the mark

Democrats should steer clear of policies that promise more than they deliver:

Income replacement over work

While some propose universal basic income (UBI) as the primary solution to AI-driven job loss, this overlooks the core goal of economic policy: increasing access to meaningful, well-paying jobs. Recently, Elon Musk publicly called for “universal HIGH INCOME“ through federal checks as the solution to AI-driven unemployment. The idea is gaining traction among many Democrats, too, in the form of an “AI dividend.”

New York House candidate Alex Bores’s proposed “AI Dividend“ would trigger direct payments to Americans once AI “meaningfully displaces” workers, funded by taxes and equity stakes in AI companies. Former Harris deputy campaign manager Rob Flaherty made a similar case in the New York Times, calling for an “AI dividend modeled on Alaska’s oil dividend” and arguing that “the basics of middle-class life can’t depend on having a traditional job.”

Both proposals treat income replacement as the answer to a question about work itself. Bores softens that premise with “insurance policy” framing, while Flaherty states it outright. Either way, the conclusion gives up too much, too early, accepting a future in which most Americans lose access to meaningful work, and the government’s role narrows to distributing the proceeds of a shrinking job market.

The case for income replacement also rests on a shaky economic premise. Flaherty’s framing also treats AI as a zero-sum transfer of wealth from workers to capital, ignoring that technologies have historically reshuffled labor rather than destroying it in aggregate. It also ignores consumer surplus: AI already provides inexpensive tutoring, coding help, medical information, and translation to people who could never have afforded those services before.

As Gene Sperling argues, replacing work with income support neglects the need to improve wages, job quality, and mobility, potentially accepting a future with limited opportunities for contribution. Although transition-based income support is helpful, the main focus must be on actively expanding and shaping job opportunities instead of substituting for them.

The robot tax trap

Another approach gaining traction is to tax AI or automation directly. Sen. Sanders (I-VT) has long backed a “robot tax” on companies that replace workers with automation, and Bores’s and Flaherty’s dividend proposals share the same basic premise: fund payouts through taxes on AI companies. But that same logic would have called for taxing tractors, washing machines, ATMs, and self-checkout kiosks. Each displaced jobs. Each also made the economy more productive and created new ones.

Leaning on an “AI tax” to fund healthcare, housing, and childcare deflects from the real reasons those sectors are broken: restrictive zoning, provider consolidation, licensing barriers, etc. These problems long predate AI and require direct supply-side fixes.

Research on automation taxation finds the same pattern: higher costs, firms relocating to friendlier jurisdictions, slower productivity growth, and no clear gains for workers. The better move is to shape where AI gets deployed, steering it toward uses that support workers rather than replace them.

Retraining without a destination

Standalone retraining programs, disconnected from real hiring demand, have a persistently weak track record. Programs that measure success by enrollment instead of employment, lack employer coalitions, or fail to track wage outcomes tend to produce weak results. This is a fixable problem, and aligning training with real jobs is one of the highest-impact changes policymakers can make.

The risks of getting AI policy wrong are clear. Poorly designed policies could allow AI to deepen insecurity, widen inequality, and slow innovation, all without making workers meaningfully better off. That would be the worst of both worlds.

Democrats have a real chance to shape how AI transforms work, but only if we resist the temptation of easy answers. The goal shouldn’t be to replace work with checks or to slow progress with taxes. It’s to build an economy where more people can contribute, adapt, and get ahead. That means doubling down on what works, and having the discipline to walk away from what doesn’t.

Hope Ledford is the Director of Civic Innovation Policy at Chamber of Progress, leading work on autonomous vehicles, the gig and sharing economies, workforce issues, delivery, public safety and parking technology, and telemedicine.