Steering AI’s Economic Impacts, Before They Arrive

AI can radically disrupt working lives and material well-being, for better or worse. The same technology that could accelerate the discovery of new medicines and raise farmer yields in developing countries is also concentrating wealth and widening inequality between countries. Which future we end up with, and for whom, is not fixed. It will be determined by the choices civil society, labor, industry and policymakers worldwide make in the next few years.

That uncertainty is visible in the data coming out right now. U.S. employers have cited AI as the leading reason for job cuts every month this year, tying more than 100,000 layoffs directly to AI through June, nearly double the total for all of 2025. But it’s contested whether this reflects a genuine shift in what AI can do to knowledge work, or “AI-washing,” where companies use AI as convenient cover for a correction from pandemic-era overhiring.

PAI’s Guidelines for AI and Shared Prosperity, released in 2023, were used as a north star for the U.S. Department of Labor’s 2024 principles to advance equity and job quality. But since then, uncertainty about the future path of the economy has continued to grow. This is because AI’s potential impact on the labor market depends on the pace of technical progress in areas like post-deployment learning and robotics, how quickly industries adopt and see returns, and the policy environment that shapes both.

Given the high uncertainty about AI’s technological progress and adoption, scenario planning is a powerful tool for strategic planning and preparation. But given the multitude of factors that could affect AI’s trajectory, effective scenario planning requires diverse expertise, bringing those who understand AI’s technical trajectory together with those who understand how worker power can shape technology.

To address this need, earlier this year, PAI launched a new initiative – Shaping Economic Futures in the AI Era and convened a multi-stakeholder Labor and Economy Steering Committee to guide the effort. This work is designed to leverage scenario planning as a tool to help society shape what future we want. Scenarios are tools for pressure-testing assumptions and surfacing the choices that matter most, giving participants a shared, concrete future to argue over rather than an abstract debate about probabilities.

As part of this effort, on July 29–30, we’re co-convening a scenario planning workshop in Washington, D.C. with the Windfall Trust, bringing together leaders from labor unions, industry, and civil society alongside our Labor and Economy Steering Committee. Participants will spend the workshop inhabiting two stylized 2030 scenarios, and then work backward to what we’d need to start doing in 2026 to steer toward better outcomes under either one. These two scenarios draw in part on the OECD’s February 2026 report modeling a range of AI capability trajectories through 2030:

In the ‘Slow’ scenario, AI keeps improving but at a decelerating pace, remaining concentrated in knowledge work. Underemployment, contract work, and income volatility spread beneath the surface, without the headline unemployment numbers that would otherwise prompt a policy response.

In the ‘Fast’ scenario, job losses outpace new job creation and automation gains concentrate rapidly among asset owners, arriving far faster than tax, safety-net or fiscal systems built for gradual change can adapt.

The workshop is held under the Chatham House Rule so participants can speak candidly about where they see risk and opportunity. Insights from the convening will be used to update PAI’s Guidelines for AI and Shared Prosperity and develop Recommendations robust to multiple scenarios.

Scenarios are a tool and a starting point, not the destination. Alongside them, we’re bringing a starter list of draft recommendations spanning areas like international coordination, public investment, taxation and market structure, and worker power—building on our 2023 Shared Prosperity Guidelines and initial discussions with our Labor and Economy Steering Committee this spring. We’re asking participants to pressure-test that list, flag what’s missing, and help us identify the “no-regrets” actions worth taking regardless of which future could unfold. Workshop discussions will help shape a detailed report about the Scenarios by the Windfall Trust developed in partnership with PAI this Fall, and inform Recommendations that PAI will develop by year-end leveraging this scenario analysis. We don’t think the uncertainty over AI’s pace is a reason to wait for clarity that may not arrive in time. It’s the reason to plan for more than one future at once, and to start building the actions now that would hold up under either.