Policymakers recently introduced two bills to address AI’s impact on the workforce. First, the AI-Related Job Impacts Clarity Act, cosponsored by Senators Hawley (R-Mo.) and Warner (D-Va.). Second, the AI Workforce PREPARE Act from Senator Jim Banks (R-Ind.), with cosponsors Senators Maggie Hassan (D-N.H.), John Hickenlooper (D-Colo.), and Jon Husted (R-Ohio). It is encouraging to see several lawmakers take up this critical issue; however, only one of the bills fully grasps how AI is actually reshaping the labor market and where those shifts will show up in the data. 

Predictions from AI company executives warning of large-scale workforce disruption, coupled with headlines linking AI to mass layoffs, have helped make the issue of AI-driven job displacement a potent political force. Addressing this widespread anxiety is a bipartisan issue, with both Republican and Democratic lawmakers highlighting the protection of the American worker as a top priority. 

Both of these new bills attempt to provide policymakers with more actionable information about AI’s impact on the workforce. But getting this right requires designing data collection efforts around an accurate understanding of how businesses actually make decisions about their workforce and deploying new technologies like AI. 

How the AI Workforce PREPARE Act Gets it Right

To this end, the AI Workforce PREPARE Act hits the nail on the head. The bill combines an ambitious plan for action with a realistic understanding of how AI affects work. The Act aims to enhance federal agencies’ capacity to assess AI’s impact on the workforce and to make education and job-training programs more effective. It introduces several key provisions, including the creation of an AI Workforce Research Hub to support implementation of the AI Action Plan and the establishment of prize competitions to better understand AI’s impact on work. 

What sets the AI Workforce PREPARE Act apart, however, is its clear and nuanced understanding of how AI is shaping the labor market. That grounded starting point informs what the bill seeks to measure: three key signals that, taken together, offer a credible path forward for understanding AI’s impact on work.

Three Signals That Matter

First, the AI Workforce PREPARE Act targets how AI is being adopted and used within firms, from encouraging model providers to share anonymized data on workplace adoption to updating federal surveys to capture which AI tools workers are using and how intensively. Understanding how AI is used within firms provides policymakers with a grounded picture of AI's impacts on workers. For instance, if a sector uses AI to automate a large share of tasks rather than augment them, this likely signals downward pressure on employment or wages.

Second, the bill seeks to capture another key labor market signal: which occupations are most exposed to AI. It directs the government to identify a set of “AI-sensitive” occupations and track how AI model advances shift demand for workers in those occupations. This gives policymakers an early indication of where pressures may emerge or which jobs may require more targeted support. 

Finally, the bill would attempt to track the labor market directly. It calls for new data on job-to-job flows – how workers transition between roles and industries – which will become especially important as AI reshapes demand for certain types of work. Job-to-job flows offer an early and nuanced indicator of how workers adjust when labor demand changes – whether they find comparable work, trade up or down in wages, or leave their field entirely. This offers a more comprehensive view of how AI is reshaping opportunities across the economy, not just within individual firms. 

Where the AI-Related Job Impacts Clarity Act Misses the Mark

While the AI Workforce PREPARE Act aligns measurement with how AI’s labor market impacts unfold in practice, the AI-Related Job Impacts Clarity Act overrelies on easily distorted metrics. The bill tasks federal agencies and companies to report at least four distinct categories of data related to AI-driven automation: the number of individuals laid off due to the use of AI; individuals hired due to the incorporation of AI; positions not filled due to the use of AI; and individuals being retrained due to AI.
In doing so, the bill overlooks the powerful incentives companies may have to both under- and over-report AI’s impacts on their workforce. In some cases, public reporting will strongly incentivize some firms to downplay AI’s effects to avoid reputational or ethical backlash, especially as layoffs deemed “AI-related” become an increasingly politically charged issue. And in others, firms have incentives to over-report AI’s impacts, given that the market rewards firms that attribute cost-cutting or restructuring to AI. 

It is important to note that both bills include a requirement to report when AI is involved in mass layoffs. To be clear, asking companies to report AI-related workforce changes can yield useful information. But given the incentives involved, this is most useful when these are treated as part of a broader array of useful signals rather than the foundation of measuring the impacts of AI on the workforce. And, while the broader signals captured by the AI Workforce Prepare Act would also be subject to some pressures around misreporting, they are more structural indicators of AI’s impact on the labor market and thus likely to be more resistant to gaming than self-reported attribution. 

Critically, the bill is also based on a misunderstanding of how AI interacts with the labor market. AI is undeniably reshaping work, contributing to layoffs, unfilled positions, and the creation of new positions. But not in the straightforward, job-by-job way that the AI-Related Job Impacts Clarity Act implies. Instead of displacing workers outright, AI typically diffuses into workflows, taking over some tasks, altering or augmenting others, and blurring the line of where human labor ends and automated labor begins. 

This means that, rather than a robot replacing a factory worker in a single, clear, and easily measurable step, AI-driven automation unfolds more indirectly. An accounting department might introduce generative AI tools into auditing and reporting workflows, fully automating some tasks, such as spreadsheet generation or data cleanup, while using AI to support others, like drafting reports. Over time, the department keeps what works and discards what doesn’t. Team structures and employee headcount gradually shift in response, but these amorphous changes don’t fit neatly into monthly classifications under categories such as “AI-related layoffs” or “positions not backfilled.” While it is encouraging to see policymakers seek better visibility into AI-related layoffs, grounding that approach in unrealistic classifications undermines their efforts.

The Stakes of Measurement

Workforce policy is only as good as the data that it is built on. How Congress chooses to measure these changes will shape how well it can respond to them. To this end, the AI-Related Job Impacts Clarity Act falls short. The AI Workforce PREPARE Act, however, offers a more promising way forward. It shifts away from relying on unreliable data and instead focuses on the underlying patterns that are more likely to reveal workforce outcomes. In doing so, it offers policymakers a much more grounded foundation for understanding – and responding to – how AI is actually impacting the labor market.