AI Job Loss and Basic Income: The Gap Wage Insurance Cannot Fill

RAISE US lists wage-insurance pilots among the experiments it plans to fund with states and researchers to prepare for AI disruption. Its government agenda names partnerships in Arkansas, Maryland, Utah and Connecticut. These are the organization's stated plans. [1]

I welcome testing income protection. But my first test is whether help reaches someone whose applications produce no offer. A payment that depends on finding another job cannot, by itself, protect people while that search fails.

That is why I take basic income seriously, while refusing to treat universal prosperity as AI's automatic destination.

Wage insurance helps once there is a wage

Under wage insurance, a displaced worker receives support after returning to employment at lower pay. A May 2024 New York Fed staff report examined that approach within the U.S. Trade Adjustment Assistance program. Its summary reports that eligibility increased employment in the short run and cumulative earnings over the longer run, largely because workers spent less time without employment. [2]

This deserves a fair hearing. My interpretation is that cushioning a pay cut can make taking another job more viable. The findings also challenge the assumption that accepting lower pay necessarily means lower total earnings over the years that follow.

But cumulative earnings can improve because someone works more months; that does not establish recovery of their previous wage. Evidence from trade displacement also cannot establish how the policy would perform if AI reduced hiring across many occupations simultaneously.

I would support testing wage insurance as one component of a broader system. Its employment condition is exactly why complementary protection matters.

Reemployment can still leave a loss

An August 27, 2026, Bureau of Labor Statistics report illustrates the different outcomes after displacement. Among people displaced during 2023–2025 from jobs held for at least three years, 66.1% were reemployed in January 2026, 18.3% were unemployed and 15.7% were outside the labor force. These are historical displacement figures; they do not attribute losses to AI. [3]

Among the narrower group who lost full-time wage-and-salary jobs, returned to full-time wage-and-salary work and reported the relevant earnings, about 49% earned at least as much as before. [3]

Those findings describe different problems. Someone outside the labor force is not necessarily searching unsuccessfully. Someone back at work may still have lost income. I think declaring the transition complete once a worker is hired sets an inadequate standard.

For policy design, I would distinguish protection while someone has no earnings from protection against a lower wage after reemployment. Improving the second leaves the first unresolved.

Basic income changes who can receive help

By universal basic income, I mean a regular payment to everyone in the covered population, independent of employment. Its appeal here is straightforward: support could continue while employers decline to hire you.

OpenResearch's randomized trial offers relevant, limited evidence. During 2020–2023, 1,000 adults received $1,000 monthly and 2,000 comparison participants received $50. Participants were ages 21–40 in Illinois and Texas, with household incomes no higher than three times the federal poverty level. [4]

The organization reports that household income was higher when transfers were included, but lower when they were excluded. Recipients averaged 1.3 fewer hours of work per week and were two percentage points less likely to be employed than comparison participants. Employment nevertheless rose in both groups during the study. [4]

I would take both findings seriously. The transfers increased household resources, and recipients did somewhat less paid work. Income security and maximizing hours worked are different objectives. In my judgment, giving people some freedom to work less can be a legitimate benefit; the employment response also belongs in any assessment of national costs and funding.

This temporary, targeted experiment did not test a permanent universal program or an economy where AI had made jobs broadly unavailable. [4] I regard it as evidence that cash can change people's choices and resources, with substantial questions remaining about scale and permanence.

Universal high income is a bigger promise

I see three separate ambitions: preventing destitution, replacing a lost salary and making everyone prosperous. I would keep those promises separate when evaluating an AI income proposal.

Universal high income would need an explanation of who receives the proceeds of AI production, how payments are funded and what those payments can buy. More capable software does not itself specify those arrangements. I would regard a prosperity promise without them as speculation.

That leaves room for ambition. A society could choose to distribute technological gains broadly. My objection is to presenting that choice as something technology automatically makes for us.

I also reject the claim that inevitable mass poverty and chaos are the only alternative. The evidence here establishes neither universal abundance nor unavoidable collapse. My concern is more immediate: policymakers could spend years debating eventual prosperity while people face income losses that their proposed protections leave uncovered.

Cover the unsuccessful search

My preference is for support during unemployment alongside help with earnings losses after reemployment. Basic income is one possible foundation; targeted support could also be designed to reach people without a new job. These studies do not establish which national design would work best.

The RAISE US government's page supplies no pilot payment amounts, eligibility rules or application timetable. [1] For career planning, I would treat access as unconfirmed until published terms establish who can participate. An announced experiment should not become the reason to pause a job search or commit to lengthy retraining.

I expect the appropriate balance to depend partly on how much paid work remains available. If opportunities expand, wage insurance could help people move into them. If hiring contracts broadly, protection tied to reemployment would become less useful precisely when more people need income support. That would be a serious failure of coverage, even without universal unemployment.

I want protection that still works after a worker does everything reasonably asked of them and receives no offer. The absence of a hiring decision should never be mistaken for the absence of need.