How Americans Feel About a World Without Paid Work
Exclusive new polling about our AI future
Imagine a world where robots and AI do almost all paid work. The folks at Blue Rose Research asked 12,000 Americans a series of questions about what they would want in that scenario, and shared their results exclusively with The Bully Pulpit. While such a world is not imminent — and may never happen at all — the survey results provide fascinating insights into how Americans think about AI, the benefits of work, and the essentials of a fair society.
I have two main takeaways from the polling data:
1. Americans believe that gains from AI should be broadly and equally shared
Americans have a bit of a socialist streak when it comes to AI. When asked who should own robots and AI systems, a plurality says that every citizen should own an equal share:
An even larger percentage of Americans believe that the profits of AI companies should be shared with the broader public:
There is some—but not much—partisan divide in these results. Trump voters are somewhat more likely to support private ownership than Biden 2020/Harris 2024 voters. But even among Trump voters, support for broad sharing of AI products and profits outstrips private ownership. And there is surprisingly little disparity across age groups.
These survey questions don’t test specific policy ideas but the thrust of the answers is clear: Americans want the gains from AI to be broadly and fairly shared.
2. Americans value work—and therefore prefer a “jobs guarantee” to a universal basic income as a response to job loss
You might think Americans would welcome a world in which AI takes care of most work and they could just do what interests them. That is not so:
There is a bit of variation by age and between people who voted in 2024 for either major presidential candidate and those who didn’t. Both younger Americans and non-voters are more likely to think that being able to pursue their interests is a good thing. But even among those groups it is less than 30%. There is robust overall support for the notion that jobs provide dignity and purpose.
Not surprisingly, that translates into a preference for the government providing support for people via a jobs guarantee rather than a universal basic income untethered to any work:
Again, there are some variations across groups. Younger and older Americans are more likely to prefer a jobs guarantee, while middle-aged Americans are more evenly split (perhaps because they are more likely to be working!). And consistent Democrats (those who voted for both Biden in 2020 and Harris in 2024) are more inclined to UBI while Trump voters, swing voters, and nonvoters prefer a jobs guarantee.
The preference for a jobs guarantee-type policy comes through even more strongly in responses to similar questions:
My bottom line from these findings (as well as from the projections about the potential economic impacts of widespread AI adoption, which I wrote about here): policymakers should spend more time exploring what a workable jobs guarantee program could look like. If AI does in fact replace millions of jobs without creating millions of new ones, then a jobs guarantee funneling newly unemployed people into societally productive work may be better than Trade Adjustment Assistance-like retraining programs that don’t have the best track record of success and are even more unlikely to work if there are simply fewer available private-sector jobs.
As I noted at the beginning of the piece, we may never get to a world in which AI replaces most paid labor. But I find these results fascinating because they tell us something about how the American public views the world. Most want the dignity and sense of purpose that comes with a paid job, even if it’s not required to receive a sustainable income. Most want AI tools and any AI-related profits to be broadly and equitably shared.
The Trump Administration is ignoring the public’s views entirely. They are on the verge of issuing an Executive Order attempting to preempt the states from regulating AI—disregarding the public’s concerns about AI safety and job loss to cater to the industry’s desire for a light-touch regulatory approach. This creates an opening for Democrats to take the public’s side on an issue that is already growing in salience and could end up being one of the public’s top concerns in the 2026 and 2028 elections.
Great piece. I’ll add that shorter workweeks (4-day, 32 hours) are one big policy idea that’s often missing from the discussion of securing a just transition for the working class. If there’s massive job displacement from AI, shorter workweeks help spread the available human labor around, likely more effectively than a jobs guarantee program would. We saw how the US shifted towards shorter workweeks in the 30s/40s following the job loss due to the Great Depression. Could turn to that model again in the face of such massive workforce change.
Your review would have gone further, IMHO, by pointing out that both UBI and a path to contribution have value. A person may fluctuate between those as a matter of planned and unplanned events in a lifetime. Did the survey ask users what they would want if disabled, for example? Without testing/surveying every boundary condition (and the people already living at those boundary conditions) we are stuck reimagining another world that isn’t going to work.
Further, I haven’t looked at the Trade Act retraining data yet, but a regulated path to work in this scenario sounds a bit 1984. Interests can very much equal a path to contribution. We are wired to contribute, contribution is helping. We cannot regulate contribution through interests or work. One finds both contentment and genius in the quiet moments not in structures, especially mandated ones. And yet both contentment and genius lead to deep contribution and a reasonable world.
Something is off about this data.