Advancing Economic Participation in the AI Era: A Private- and Public-Sector Agenda
Advances in artificial intelligence (AI) are poised to reshape the US economy at an unprecedented scale and speed. In Advancing Economic Participation in the AI Era: A Private- and Public-Sector Agenda, Glenn Hubbard and Luke Pardue argue that while AI will likely drive higher aggregate income and wealth in the long term, the gains are likely to be unevenly distributed, and the path to this end state will involve a disruptive transition. A failure to pay attention to this transition may generate a backlash that puts those long-term gains at risk. Responding to this shift will require robust action from the private and public sectors around social insurance, workforce development, and tax reform.
Hubbard and Pardue lay out several principles to guide the private- and public-sector reform agenda:
- AI progress will be fast and uneven. AI-driven disruption is already unfolding with disparate consequences for workers in different occupations and industries. The high level of uncertainty surrounding AI’s impact on the economy calls for a risk-management approach that is flexible to different forms of disruption.
- Maintaining connection to productive activity is critical to a flourishing society. Work not only generates earnings but is also a source of identity and purpose. Employment has been associated with higher overall well-being, and economic disconnection generates large personal and social costs. The authors argue that responses to AI developments should prioritize active participation in the economy among working-age adults.
- Building trust in public institutions is crucial. The AI revolution is occurring at a time when trust in the country’s major public institutions is at a historic low. Navigating this challenge will require leveraging well-established institutions, such as America’s community college system, while also experimenting with new approaches.
- Long-run outcomes should be distinguished from transition outcomes. Focusing too heavily on long-run, “steady-state” outcomes risks ignoring that the economic transition will likely be disruptive for some workers. If individuals do not believe AI advances will benefit them, they may take steps to slow the diffusion of the technology.
In the long term, AI’s impact on employment will depend on two factors: whether it automates tasks without creating new ones that rely on human expertise and whether productivity-enhancing technology increases demand for a sector’s output. Under these conditions, four scenarios could emerge:
Scenario 1: AI automates enough tasks to reduce labor demand. It is likely that AI tools can perform enough tasks within certain occupations to reduce labor demand. Such jobs may involve routine, programmable tasks, mirroring the effects of prior episodes of technological disruption, but AI tools may also reduce demand for certain jobs involving nonroutine, cognitive tasks.
Scenario 2: AI reshapes work but keeps “humans in the loop.” In certain occupations, AI technology may automate many but not all tasks, and the key factor in determining how AI will affect wages and employment within these reshaped jobs is the type of tasks humans are left to perform. In many cases, AI may automate the rote tasks within knowledge work, leaving those that require expertise, context, and judgment, such as evaluating output from AI tools or emphasizing the human-centered components of the occupation.
Scenario 3: Economic value flows to “bottleneck jobs.” Value shifts toward tasks AI cannot yet perform, particularly in the “relational sector,” where the human element is a critical part of the output. The economic value of jobs with those tasks will increase.
Scenario 4: AI creates demand for new expertise. Increases in AI use will create demand for new occupations entirely, as was the case in prior episodes of major technological change.
How firms choose to adopt AI will shape which scenario prevails. The political economy risks during the transition period should push business leaders to recognize the value of being good stewards of their current workforce and of the communities in which they operate. They should develop firm-level strategies to incorporate AI tools in a way that complements tasks within human work rather than replacing whole occupations. Over the long term, firms should invest in worker training to ensure their skills and expertise keep up with these reshaped jobs.
Because the labor market risks posed by AI are too large for individual parties to bear, Hubbard and Pardue call for a modernized social insurance program designed to temporarily mitigate income loss and reconnect workers to new opportunities. Such an effort will involve expanded wage insurance to speed reemployment for older workers, Personal Reemployment Accounts that provide flexible funds for workers to use on training services along with a reemployment bonus, and a modernized workforce development system that strengthens ties between employers and training providers.
Finally, Hubbard and Pardue propose reforms to the US tax structure that can raise revenue amid shifting forms of economic activity while encouraging domestic savings, investment, and growth. They recommend a transition away from the current income tax structure and toward a consumption tax by pairing a business cash-flow tax with a progressive wage tax. This structure will tax firms’ excess profits and remove distortions of saving and investment decisions that are present in the current system, while maintaining the progressivity and pro-work elements of today’s tax code.
Suggested Citation: Hubbard, Glenn and Pardue, Luke. 2026. “Advancing Economic Participation in the AI Era: A Private- and Public-Sector Agenda.” In The American Economy in a New Era, edited by Melissa S. Kearney and Luke Pardue. Washington, DC: Aspen Institute.