Artificial Intelligence and Robotics: From a Labour and Tax Perspective


May 2018

Modern information technology, intelligent algorithms and production robots are strongly influencing the working world in the 21st century. Everyday tasks are already being performed by intelligent algorithms. The question thus arises as to what the future world of work will look like. This article will give an overview of the future labour market as well as the impact of artificial intelligence (AI) on employment and labour law as well as tax issues. 

Cost savings thanks to robotics

Particularly in the industrial sectors in Western high-wage countries, automation and the use of production robots lead to considerable savings in the cost of labour.

While one working hour in the production plants of the European automotive industry costs more than EUR 40, the costs for the use of a robot range from EUR 5 to EUR 8 per hour.[1] Today, a production robot is already cheaper than a worker in China.[2] The ODI’s report on "Digitalisation and the future of manufacturing in Africa" concludes that by 2034 and 2038, production and 3D prints in the US will be cheaper than producing in Kenya or Ethiopia thanks to AI and robotics.[3]

The future labour market

A first study by Frey/Osborne published in 2013 assumes that 47 % of American jobs are at a high risk of being automated.[4] According to an OECD report from 2017, 14 % of jobs in the 32 member states are highly automatable. An additional 32 % of the jobs are likely to change significantly.[5] The 2017 McKinsey report reached a similar conclusion: 15 % of the global workforce is subject to a high risk of automation.[6]

However, all reports state that significant differences between countries and sectors are possible. In general, the OECD report reasons that jobs in the Anglo-Saxon and Nordic countries are less automatable than jobs in eastern or central Europe. This is because automation affects jobs in the manufacturing industry first. Countries with a high number of production sites and a low degree of automation are therefore affected the most.

Nevertheless, both blue- and white-collar sectors will be equally affected. Certain types of jobs will disappear partly or completely. At the same time, it is expected that new jobs will be created in the service sector. These range from data detectives and AI business development managers to cyber city analysts and AI-assisted healthcare technicians.[7]

Especially in the medium-wage sector, routine jobs will be eliminated. One-third of the jobs that require a bachelor's degree could be replaced by intelligent software in the coming years. Even well-paid and high-skilled jobs like those of stock exchange traders can be rationalised. Goldman Sachs, for example, reduced its traders from 600 in 2014 to 2 in 2017.[8] However, jobs will not be lost abruptly, instead a gradual transition will take place, which has already commenced.[9]

Low-wage countries such as China, India and Bangladesh are currently still benefiting from their surplus of low-skilled workers. The wage costs are lower than the costs for the acquisition, development and maintenance of digitalised equipment.[10]

But it is very likely that companies located in low-wage countries will soon invest in the relevant systems to improve their productivity and competitiveness in the long run.[11] Some companies, like Adidas, have decided to produce in their countries of origin, using production robots with only a few workers in their "speed factories". Thus, soon low labour costs will not be a competitive factor for a country anymore. In the long term, Western and Southeast Asian countries will be considered the winners of digitalisation because of their technological head start.[12]

The role of human workers

The role of human workers is changing. Due to the enhanced use of AI, new challenges are in store for employees and unions.

Unions will work to ensure that the introduction of AI does not cause mass dismissals. The reintegration of untrained workers into a digital labour market will be one of the most difficult challenges. To avoid this, unions will advocate further training, advanced training and retraining of employees to shift as many jobs as possible to other sectors.[13]

In addition, they will use their influence to avoid some technical developments that, in the opinion of some unions, could lead to health risks (e.g. employees' permanent availability), social dumping (e.g. platform industry) and total surveillance (e.g. wearables).

They also will potentially expand their constituency by representing the increasing number of freelancers in the gig economy.

The growing platform economy

Instead of relocating production sites to low-wage countries or using temp agency workers, companies will focus on their core competencies and outsource other activities in a cost-effective manner via internet platforms.

It is a global trend that Work 4.0 will take place outside traditional employment structures, creating a rise in self-employment.[14] Even in European countries, where the average hourly paid wage is usually higher than the salary that can be earned through crowdworking on internet platforms, the platform economy is becoming more and more common. Larger companies are increasingly using external workers rather than hiring new employees.

The crowdworker only accepts the platform's terms and conditions. An employment relationship does not exist. Instead, there is a brokerage or service contract. This means that there is no obligation for either party to accept the other's offers, but there is a payment obligation after the completion of the agreed task.[15]

However, there are several challenges: defining crowdworking as a freelancer relationship without integrating them in one's own or the platforms' business; establishing working conditions, e.g. for compensation; finding out which tax regime and social security and welfare rules apply.

Robot tax

With the growing importance of robots and a decline in the need for human workforce in the operation of many businesses, the idea to impose a tax on robots has been developed in several countries such as the US, Switzerland and France.

This suggestion raises many questions from a tax policy perspective: while some observers see it as an incentive to keep as many workers as possible, there seems to be a contradiction between providing companies with tax incentives connected to technological investments (such as accelerated depreciation or expensing of capital expenditures) and taxing robots. Generally, one may wonder whether taxation is the most appropriate tool for drawing conclusions regarding a fundamental change in the organisation of the labour environment.

The main function of a tax on robots would probably be of a budgetary nature. Bearing in mind that many states levy taxes on employees’ wages and therefore anticipate a strong decrease of tax revenue connected to this specific base, taxing robots would help compensate for a massive loss of public revenue.

Although a “robot tax” still lacks accurate technical content, there are already many issues related to the design of such a tax. The first one is the identification of the person who should be liable to pay this tax: while the owner (either legal or economic) of the robot seems to be the obvious taxpayer, some academics go as far as to wonder whether robots could (or should) be considered as autonomous tax subjects. Would this tax be deductible from the base of other taxes? Which level of taxation would be appropriate, considering the goals pursued by this mechanism?

Many issues will also arise in the international arena. It seems clear that isolated initiatives will create distortions in international tax competition. This calls for a reflection on the way states could or should coordinate at a supranational level.

From a technical perspective, the question will also arise whether this tax should be characterised as “income tax” to the extent that the use of robots would help employers save remuneration costs. This theoretical issue is not without practical implications: if this were to be the case, then one might argue that treaties on the elimination of double taxation should be applicable where several countries intend to levy concurrent taxes on the same robots. The question whether the notional income which would be subject to tax is “income from employment” or “other income” under existing tax treaties is also a tricky one… and the answer may well dramatically affect the allocation of taxing rights between states.

Needless to say, the idea of taxing robots challenges the current concepts of both domestic and international tax law.

Fundamental concepts of law to be revisited

Robots have stimulated the imagination of human beings for several thousand years: artificial servants and companions seem to be present as far back as in the ancient Greek myths. To a great extent, these myths are now becoming reality, with the effect that fundamental concepts of law must be revisited in light of recent and rapid technological developments. It remains to be seen how quickly law will follow technology… and to what extent lawyers themselves will be replaced by robots.[16]

[1] (last accessed on October 16, 9:17 a.m.).

[2] "Kollege Roboter," Fokus No. 38/2015 of October 19, 2015, p. 69.

[3] (last accessed on April 06, 4:28 p.m.), p.3.

[4] Frey/Osborne, "The Future of Employment: How Susceptible Are Jobs to Computerisation," 2013, p. 1.

[5] Nedelkoska, L. et G. Quintini (2018), « Automation, skills use and training », Documents de travail de l'OCDE sur les questions sociales, l'emploi et les migrations, no. 202, Éditions OCDE, Paris, (last accessed on April 12, 10:57 a.m.).

[6] (last accessed on April 12, 11:04 a.m.).

[7] (last accessed on April 12, 7:44 a.m.).

[8] (last accessed on October 14, 10:17 a.m.).

[9] Brzeski/Burk, "Die Roboter kommen, Folgen der Automatisierung für den deutschen Arbeitsmarkt," 2015.

[10] (last accessed on October 13, 2:38 p.m.).

[11] (last accessed on April 12, 11:35 a.m.).

[12] (last accessed on October 13, 3:15 p.m.)

[13] (last accessed on October 15, 2:24 p.m.).

[14] "Automation and Independent Work in a Digital Economy", OECD Policy Brief on the Future of Work, May, 2016, p.3.

[15] Wisskirchen/Schwindling, "Crowdworking im Lichte des Arbeitsrechts", ZESAR, 8/2017, p. 318-327.

[16] J. Markoff, "Armies of Expensive Lawyers, Replaced by Cheaper Software", N.Y. TIMES, Mar. 5, 2011, at A1.


Dr. Gerlind Wisskirchen
Picture of Daniel Gutmann
Daniel Gutmann
Jan Schwindling
Research Associate