Taxing Robots by Xavier Oberson, Professor at Geneva University, attorney-at-law

The development of artificial intelligence (AI) and of robots in particular is very likely to have a tremendous impact on the job market and the economy as a whole. Robots are not only replacing industrial workers, but also are now more and more also used in the service sectors. Today robots, slowly, become lawyers, doctors, bankers, social workers, nurses and even entertainers. The impact of the development of AI on the economy is however controversial. The majority view maintains that, like in the past industrial revolutions, technology will foster global growth and create more jobs. On the other hand, the so-called "pessimists" hold that this time is different. Robots are not only replacing the arms of workers but the purpose of AI is eventually to replace human brain. The development of AI could also have an impact on the unequal distribution of wealth among people.

In our view, we do not know the future and cannot tell who is right or wrong between optimists and pessimists. However, it is rather probable that many workers would lose their jobs and it is also unlikely that sufficient new jobs would appear on the market for many unemployed workers. This is the reason why we think that we should consider right now some potential solution to this issue. The author considers that a tax on robots could represent an adequate solution[1]. To find a consensus on such proposal and elaborate the proper alternative requires however a consider amount of time, as the projects for taxing the digital economy, following Action 1 of the G20/OECD BEPS Initiative, have demonstrated. This is why we should act now and start analyzing the various issues raised by the introduction of a robot tax, including tax on the use of artificial intelligence.

While the effective impact on labor remains controversial among economists, solutions have to be analyzed because, should massive workplaces for humans disappear in the future, a triple negative effect could occur. First, significant tax and social security revenues would be lost because in most countries, labor income represents one of the most important tax and social security base. Second, the need for additional revenues would increase for the states in order to support the growing number of unemployed human workers. Third, because of the decrease of labor income, and human labor in general, consumption would also go further down.

Introducing a tax on robots should be justified both from an economic and constitutional standpoint. The economic justification of a robot tax, arguably, could be based on a reasoning that enterprises using robots, instead of human workers, spare the salaries that they should have paid for similar human activities. From a constitutional standpoint, a tax on robots would require that all enterprises using robots or artificial intelligence with an impact on human workers should be treated in a similar fashion. In addition, the application of the principle of ability to pay raises new and delicate issues because, so far, robots, notably “smart robots”, do not benefit as such from an ability to pay. This could however change in the future.

As a consequence, we are of the opinion that the introduction of a robot tax could be implemented but following two different stages. At a first stage, we could focus on taxing the use of robots (or artificial intelligence). At a second stage, granting a legal personality to robots could lead to the emergence of an electronic ability to pay, which should be recognized for tax purposes. After all, we have seen in the past that states, when required, could introduce new forms of legal personality..

However, introducing a tax on robots raises some complex issues, both from a domestic and international standpoint. First, a clear and practicable definition of robots would be required. Second, various ways of taxing robots should be examined. One possible solution would be to levy an income (profit) tax on the “imputed hypothetical salary” robots should receive from equivalent work done by humans. An alternative, and simpler, system would be to impose a lump-sum amount representing an approximated ability to pay the tax. Initially, this ability to pay would be attributed to the employer or owner of the robots, but as the technology evolves, the robot’s ability to pay could be recognized. Consequently, the imputed income would also become subject to social security levies. Other recent suggestions also focus on the design of a tax neutral system between robots and human workers.

Another interesting alternative is the application of the VAT on robots’ activities. Currently, robotics and automation are already subject to VAT, as part of their contribution to the value added of the production and distribution process. At a later stage, we could even consider robots as such as VAT subjects. Finally, another more traditional solution could simply be to introduce an (objective) tax on robots, similar to cars, boats or planes. Some types of these “robot taxes” have been introduced or planned in some US states on drones or self-driving cars.

Taxing robots raises issues that go beyond national borders and should be analyzed globally, considering the international taxation works at the OECD, the UN and the EU. In particular, if we were to recognize a tax capacity to robots, the proper application of tax treaties allocation and transfer pricing rules would have to be revisited.