The International Monetary Fund has suggested one way to ameliorate the impact of AI: a tax on the carbon dioxide emissions created in generating masses of energy to power the computers that many hope will do some thinking for us.
That proposal emerged on Monday in a discussion note titled “Broadening the Gains from Generative AI: The Role of Fiscal Policies.”
The thrust of the paper is that generative AI has enormous potential to improve productivity, but it could also bring less pleasant change. Automation may bring radical changes to the world of work and/or hurt national balance sheets, as fewer workers are required and tax takes drop.
The document also notes that past waves of automation have generally displaced low- and middle-skill jobs involving routine tasks, but generative AI’s capabilities are more sophisticated than past automation technologies – “potentially amplifying job losses in cognitive occupations. Consequently, the labor income share in national income may further decline, exacerbating income and wealth inequality.”
As the name of the document implies, the IMF’s boffins think fiscal policy should form part of nations’ response – along with social protection systems like social insurance, labor market programs, and social assistance schemes, labor market policies, and education and training policies.
On the tax front, the authors suggest rethinking policies on how capital is taxed. It generally attracts lower taxes than labor, but if capital uses AI to reduce demand for labor, the tax burden needs to shift.
The IMF’s analysts are also concerned about “rising market power and economic rents enjoyed by dominant firms in winner-take-all markets.” While the document doesn’t mention hyperscale operators, such outfits have already shown a capacity to spend billions on AI infrastructure – sums few can match.