Bill Gates proposes an AI token tax and Human Reserved jobs to slow automation

Bill Gates published a 6,000-word essay and gave a New York Times interview proposing a set of policies to address AI-driven labor displacement. The core ideas: a token tax on AI usage, a companion robot tax, and a new category of Human Reserved jobs that would be protected from automation. For builders, this signals a growing policy conversation that could eventually affect deployment costs, workforce planning, and where AI is encouraged versus restricted.

Gates proposes a token tax and Human Reserved jobs

Gates argues the current tax system incentivizes replacing workers with machines: businesses pay payroll taxes for human employees but can write off AI purchases as business expenses. His token tax would apply to AI usage, making automation more expensive and slowing labor displacement. The revenue would fund retraining programs and strengthen social safety nets. Gates also proposed Human Reserved jobs, roles that should remain human because workers cannot easily transition or because human interaction is essential. He cited caregiving and delivering bad medical news as examples.

What the token tax would actually do

The token tax targets the cost advantage of replacing workers with AI. If implemented, it would raise the effective price of AI tokens used in business processes, potentially shifting the ROI calculation for automation projects. Gates suggests the tax could exempt purely beneficial uses like medicine and education. The idea builds on his earlier robot tax proposal and aligns with growing anti-AI sentiment among the public and politicians.

Why builders should watch this policy direction

Even as a proposal, Gates' essay reflects a real shift in how influential figures think about AI's impact on labor. OpenAI's Sam Altman has said "entire classes of jobs will go away" and Anthropic's Dario Amodei warned AI could "wipe out" half of white-collar jobs. Gates' framework offers a concrete policy response: tax automation, protect certain jobs, and fund retraining. For builders, this could mean future regulations that affect deployment costs, especially for customer-facing automation in care, education, or construction. It also suggests that workforce retraining funding might become available, potentially creating new markets for AI upskilling tools.

The limits of a policy proposal

Gates' essay is a policy argument, not a legislative blueprint. Details on how a token tax would be administered, what counts as a token, and how to exempt beneficial uses remain unclear. The proposal also faces political and international coordination challenges. Gates has reportedly sought a meeting with Xi Jinping to discuss an international oversight body, but no concrete agreement exists. Builders should treat this as a signal of where policy thinking is heading, not an imminent change.

For now, the most useful takeaway is that the conversation around AI and labor is moving from abstract warnings to specific policy ideas. Builders who track these developments will be better positioned to adapt their deployment strategies and workforce plans.