Fiscal policy can help AI better benefit humanity

New generative AI technologies have great potential to increase productivity and improve public services, but the speed and scale of the transformation have also raised concerns about job losses and increased inequality. Given the uncertainty about the future of AI, governments should adopt a flexible approach and prepare for highly disruptive scenarios.

Fiscal policy has an important role to play in distributing the benefits and opportunities of generative artificial intelligence more equally, but this will require major upgrades to social security and tax systems around the world, a new IMF study argues.

In the face of disruptive technological changes brought about by artificial intelligence, how should countries reform their social security policies? Although artificial intelligence can eventually increase overall employment and wage levels, it may cause a large number of workers to be unemployed for a long time, resulting in a painful transition.

Past waves of automation and IMF models suggest that more generous unemployment insurance could mitigate the negative impacts of AI on workers, allowing laid-off workers to find jobs that better match their skills. Most countries have much to do in terms of expanding unemployment insurance coverage and making it more generous, improving the portability of insurance benefits, and considering the form of wage insurance.

At the same time, industry-based training activities, apprenticeship arrangements, upskilling and retraining programs can play a greater role in preparing workers for the AI ​​era. Automation or industry disappearance will leave some workers facing long-term unemployment or a decline in local labor demand - these workers will need comprehensive social assistance programs.

To be sure, the impact of AI on emerging market and developing economies will be very different, and for that reason, policymakers in these economies should respond differently. Workers in these countries are less exposed to the risks of AI, but given the size of their economies’ informal sectors, their formal social protections, such as unemployment insurance, are also relatively weak. Innovative approaches using digital technologies can help these countries expand the coverage of social assistance programs.

Should AI be taxed to mitigate labor market disruptions and pay for the workers it affects? Faced with similar concerns, some have suggested a robot tax to discourage companies from replacing workers with robots.

But taxing AI is a bad idea. Your AI chatbot or co-pilot can’t pay such a tax—only humans can. Creating a specific tax on AI could stifle productivity gains by slowing investment and innovation. It would also be difficult to implement and, if misplaced, could do more harm than good.

So what can be done to adjust tax policy in the age of AI? In recent decades, some developed countries have expanded corporate tax breaks for software and computer hardware to promote innovation. But these incentives also tend to encourage companies to automate and lay off workers. Inefficient corporate tax systems that tend to eliminate human jobs quickly should be reconsidered because they may amplify the layoffs caused by AI.

Corporate tax systems in many emerging markets and developing countries often hinder automation. In their own way, they create distortions, discourage investment, and prevent these countries from catching up in the global AI economy.

How should governments design redistributive taxation to offset rising inequality caused by AI? Like other types of innovation, generative AI is likely to lead to rising income inequality and wealth concentration. Taxation of capital income should therefore be strengthened to protect the tax base from a further decline in the share of labor income, while offsetting rising wealth inequality. This is critical because more public revenue will be needed to increase investments in education and social spending and to expand the benefits of AI.

Since the 1980s, taxes on capital income have fallen steadily in advanced economies, while taxes on labor income have risen.

To reverse this trend, strengthening corporate income tax would help. More than 140 countries have agreed on a global minimum tax rate, setting a minimum effective tax rate of 15% for multinational companies - a step in the right direction. Other measures could include: a complementary tax on excess profits, strengthening capital gains taxes, and improving tax enforcement.

The latest breakthroughs in AI reflect years of investment in basic research, including from projects funded by the public sector. Likewise, the decisions policymakers make today will shape the development of AI in the coming decades. It is imperative that all parties ensure that AI is used for the broad benefit of society and that it is used to improve education, health, government services and more. Given the global impact of this powerful new technology, it will be more important than ever for countries to cooperate.

—This blog is based on a staff discussion paper, to which Fernanda Brollo, Daniel Garcia-Macia, Tibor Hanappi, Li Liu, and Anh Dinh Minh Nguyen contributed.

Source: IMF