If artificial intelligence and automation begin to replace human labour at scale, the key economic question won’t be the speed at which jobs vanish — it will be who pays the bill. As machines generate an ever-larger share of economic value, governments will be forced to confront how to fund retraining and unemployment support in a world in which fewer people are needed to produce more wealth.
AI can supercharge inequality — unless the public learns to control it
Two priorities stand out as governments start preparing for an AI-driven shift: widening the tax base beyond labour and strengthening the institutions that help workers to adapt to technological change.
Currently, in most countries that belong to the Organisation for Economic Co-operation and Development (OECD), labour is heavily taxed: income tax and social-security contributions make up about one-third of employment costs and generate almost half of tax revenues (see go.nature.com/4ogzsr4). Technology, by contrast, is lightly taxed — and frequently subsidized — to promote its adoption and competitiveness.
Proposals to tax technology, such as a ‘robot tax’ that targets the ownership or operation of robotic equipment, have struggled to gain political support. If automation displaces millions of workers, a fiscal system that has been built to tax people, not machines, will be exposed. Public revenues will shrink just as demand for retraining and welfare spending rises.
Because AI can perform a wide range of cognitive tasks, its impact on jobs will be more far-reaching than that of earlier technologies such as computers or the Internet. When combined with robotics, AI also extends automation into manual work, such as picking and sorting goods in warehouses, quality inspection and delivery.
AI does not need to eliminate most jobs to create a policy crisis. The jobs it creates might require different skills and experience, be concentrated in certain regions or emerge more slowly than the pace at which older jobs disappear. Even if the most disruptive scenarios do not materialize, such mismatches are likely, giving governments reason to prepare now for a difficult adjustment period ahead.
Why AI hasn’t caused a job apocalypse — so far
How might governments raise revenues differently? There is a strong case for taxing capital more heavily, particularly when profits reflect market dominance or windfall gains. Much of the AI revolution relies on public inputs — from taxpayer-funded research and telecommunications infrastructure to vast amounts of user data. Allowing these gains to accrue entirely to private shareholders risks leaving taxpayers to bear the costs while investors capture the gains.