Bridgewater Estimates AI Could Displace 18% of U.S. Jobs
The firm proposes a 35% tax on AI tokens, with projected revenue used to reduce income taxes and support workers affected by displacement.
Artificial intelligence could displace 18% of current U.S. jobs over the next five years, based on an estimate tied to tasks most vulnerable to automation and a proposal for taxing machine labor.
Bridgewater Associates has proposed a 35% tax on AI tokens, defining the term as the output of AI systems and a proxy for machine labor. The firm estimates the tax could raise roughly $150 billion in 2027 and $600 billion by 2030.
The proposal comes as policymakers weigh how automation could affect employment and the tax base. Bridgewater argues that human labor carries wage-related taxes and regulatory costs that do not apply in the same way to machine labor, potentially encouraging employers to substitute AI for workers.
The proposed revenue could be used to reduce income taxes, fund programs for people affected by AI-related displacement and narrow differences between taxes on human and machine labor. Bridgewater also proposes using tax revenue or credits to acquire stakes in leading U.S. AI companies and distribute ownership to citizens on a proportional basis.
The 18% figure measures jobs whose underlying tasks are highly exposed to AI rather than layoffs already recorded. Greg Jensen, Bridgewater’s co-chief investment officer, said the estimate was developed by comparing tasks AI can perform with the tasks required for existing jobs.
“18% of jobs are pretty fully disruptable,” Jensen said in an Aug. 27 discussion.
The estimate does not establish a precise end date for the five-year period or provide a sector-by-sector breakdown. It also does not account for how many new jobs AI could create, leaving the net effect on employment uncertain.
Bridgewater said it is already using AI in internal labs and expects the technology to change how it operates and hires. The tax and citizen-equity proposals remain recommendations rather than enacted legislation.
The proposal adds to a broader U.S. debate over AI taxation and public ownership.
