If AI causes a sustained surge in the economy's productivity potential, it will significantly ameliorate America's dire fiscal picture — though by how much depends on how bumpy the path is for workers and how much the government does to assist people whose jobs are displaced.

The big picture: That's the upshot of new modeling from the Budget Lab at Yale, which shows that in the most optimistic scenario — one with strong productivity growth but without mass unemployment — the national debt would level off as a share of the economy.

• In other scenarios, where there are major job losses and the government takes steps to help those who can no longer find work, the impact on the fiscal outlook remains positive, but less so.
• Those aren't predictions or forecasts from the Yale economists, but rather the results you get from plugging productivity predictions coming from economists and technologists into a simple model.

Zoom in: The Yale team worked off of a recent paper that surveyed economists, technologists and others about their predictions for the impact of AI on GDP, labor force participation and other variables.

• They modeled a scenario from those forecasts in which productivity rises 2.5% a year over the next five years — a level that's not without historical precedent but would be a big step up from the 1.8% seen over the last decade.
• They then plugged that into a "small macro model," a nifty new tool that allows non-economists to do back-of-the-envelope modeling of how different developments would affect the economy.

Zoom out: In the Goldilocks scenario, that sustained higher productivity growth occurs amid continued full employment — essentially, labor reallocates to the AI-driven jobs of the future in a relatively gradual way.

• If that were to happen, the deficit would shrink relative to the economy — to 3.7% of GDP in 2035, compared with 6.2% in the economists' baseline.
• That would essentially stabilize the size of the national debt relative to the economy, leaving it at 100.3% in 2035, or about the same as today.

Reality check: Those assumptions about the labor market are not completely implausible — rapid, technology-driven productivity growth amid full employment happened in the 1960s and late 1990s — but neither are they the base case for leading thinkers around AI.

• They think the technology will likely create some sustained worker displacement, resulting in a significant decrease in labor force participation.

By the numbers: If that downshift occurs, it takes some of the shine off the AI-driven improvement to the fiscal outlook.

• In a scenario where the labor force participation contracts and federal assistance to those workers is on the order of current unemployment benefits — $5,500 per year — the debt-to-GDP ratio continues rising to around 108% in 2035.
• If the government jumps in with more generous help for displaced workers — on the order of the $42,000 average value of retirement benefits — the debt-to-GDP ratio rises to 112%.
• While those numbers would still prompt questions about the federal government's fiscal trajectory, they are considerably better than the 118% debt-to-GDP ratio in the Yale team's baseline with no AI productivity surge.

What they're saying: "If you're just looking at the story of increased productivity growth, it can give you an overly rosy view on how AI could affect the fiscal picture," Martha Gimbel, executive director of the Budget Lab, tells Axios.

• "There are costs that come with that productivity growth that we as a society will be expected to manage," she says.

There's something happening in the labor market: In recent weeks, key indicators suggest a reversal in a lackluster hiring trend that raised concerns about the health of the job market.

• That continued this morning, as ADP reported that the private sector added 109,000 jobs in April — the best gain since January 2025.

What they're saying: "We have seen a step-level change in the labor market," ADP chief economist Nela Richardson told reporters. "We are starting to see sectors that have been underperforming perform."

• Education and health services remained the key jobs growth drivers, but notably, other sectors rebounded — including manufacturing, which showed a slight gain in employment for the first time in two years, according to ADP's data.

The intrigue: The Bank of America Institute said its measure of employment activity rose by 1.9% in April compared with a year ago, the strongest growth in over two years.

• "The top line is that in our data, the labor market momentum — in terms of payrolls — really turned solid," BofA Institute senior economist David Tinsley told reporters.
• The institute also said that unemployment payments in customer accounts grew more slowly last month.

What to watch: According to these data sets, there is little sign that the Iran war is causing employers to hold off on hiring.

• "We're strong today, but I'm not really making any claims that this is [a] situation that will persist in the next three to six months," Tinsley said.
• "I guess the debate is, are we just in a little sweet spot between shocks — where we're getting over the 'Liberation Day' shock, but we're not really digesting the oil price shock yet? Or is this a signal about underlying momentum for the next year or so?"

The bottom line: Private estimates point to strengthening payrolls. But there are few such estimates for the unemployment rate, which the Federal Reserve is watching most closely for clues about the labor market.