Debates in the U.S. Congress over the employment impact of AI continue to intensify. A new proposal in the U.S. House of Representatives plans to impose a consumption tax on large AI companies, with the revenue directed toward creating jobs in areas such as housing construction, infrastructure, and childcare and elder care.
Tax rates increase with the unemployment rate.
The proposal was introduced by Representatives Sara Jacobs, Greg Casar, and Valerie Foushee. Under the draft, taxation would be structured in two tiers: one based on the value of tokens used by AI models, and another based on AI service revenue and certain related-party transaction income, with the higher of the two tax amounts being applied.
The draft shows that when the unemployment rate is no higher than 5%, the token tax rate starts at 2%, and the income tax rate on related earnings starts at 3%. If unemployment continues to rise, the tax rates will automatically increase. The legislators who proposed the bill argue that if Congress does not act, AI could further concentrate wealth among top corporations, while workers bear a greater share of adjustment costs.
The Senate is also advancing a similar proposal.
This proposal is also part of a recent series of AI employment legislative actions in the U.S. Congress. Previously, Foushee and Casar advocated for research into the job gains, losses, and shifts caused by AI; Jacobs also participated in another bill requiring large employers and federal agencies to report AI-related layoffs to the Department of Labor.
In the Senate, Ron Wyden proposed adjusting the tax treatment of AI data centers and establishing a new consumption tax, with part of the revenue used to support workers affected by AI. Elizabeth Warren also advocates taxing AI companies based on the energy consumption of their data centers and directing the proceeds toward worker support programs.
The tech community is also discussing the allocation mechanism.
Bernie Sanders previously proposed a more aggressive plan. He warned that AI could displace tens of millions of jobs and advocated for a one-time 50% tax on OpenAI, Anthropic, and xAI, while also enabling American citizens to share in the profits of the AI industry through equity arrangements.
In addition to direct taxation, Congress is also discussing other buffering measures. The AI Workforce PREPARE Act, proposed by lawmakers from both parties, requires the federal government to more systematically track job losses caused by AI, improve occupational forecasting related to automation, and study rapid retraining programs for the unemployed.
Calls for AI companies to bear greater employment costs do not come solely from Congress. Bill Gates has recently supported taxing AI tokens and robots, arguing that the current tax system incentivizes companies to replace human labor with machines. Dario Amodei, CEO of Anthropic, has also suggested that the federal government could consider imposing a 3% tax on revenues generated by model usage and redistributing the funds in some manner.