When AI eats the tax base

American corporate profits grew by $1.2 trillion last year, a 30% rise. Yet corporate tax revenues in the United States are falling. This paradox opens a deeper question: what happens to government finances if artificial intelligence reshapes the economy in ways that shrink the things governments tax most heavily?

The problem, as The Economist's senior economics writer Callum Williams explains, is structural. AI could create enormous value in parts of the economy that are lightly taxed, while destroying value in the heavily taxed parts. The risk is that the economy grows rapidly while governments become increasingly starved for funds.

Why labour taxes are the real target

Governments today rely overwhelmingly on taxing workers. In the United States, about three-quarters of all federal tax revenue comes from taxing labour income. European countries are even more dependent: Italy, France, and Germany tax labour particularly heavily. If AI displaces workers or reduces their wages, that enormous tax base shrinks. The US labour tax base is roughly $15–16 trillion; the entire US spending on AI is currently about $700 billion. There is simply far more taxable value in human work than in the AI economy itself.