Digitization of today’s working world is ever progressing. Almost on a daily basis, we are facing new forms of work or at least new terminology. One of the most promising but at the same time most frightening areas is artificial intelligence fueled by the notion of Hollywood blockbusters that depict world domination by machines. Even if such ideas remain pure fiction for the time being, it cannot be denied that artificial intelligence is already being put to use in numerous areas. Most people will be aware that it is not always a human being you are dealing with in call centers or computer chats. The relevance of artificial intelligence is demonstrated by the fact that the German government recently has decided to define cornerstones for an artificial intelligence strategy. These cornerstones are designed for Germany to become a global leader in research, development and application of artificial intelligence.
In everyday business life, the use of artificial intelligence is particularly evident in heavy industry and other manufacturing sectors. Where several hundred workers used to work in manufacturing plants, the number of employees has shrunk to quite a low level (for example, at the Thyssen Krupp blast furnaces). Even if politicians do not openly address this, the use and development of artificial intelligence is set to lead to the loss of certain jobs. In this respect, any advice under employment law is reduced to supporting in the staff reduction.
There are, however, several areas of employment law where the use of artificial intelligence affects the response to typical disputes in a way that is different from the past. For example, in connection with variable remuneration components, the use of artificial intelligence is an interesting topic. A classic variable remuneration component – such as a performance bonus – generally requires certain pre-defined individual targets to be achieved. The degree of target achievement is quantified at the end of the relevant performance period on which basis the decision is made whether or not a bonus will be paid. But how is artificial intelligence supposed to affect this? It is important to know that artificial intelligence is generally understood in a way that a person’s decision structures are to be reproduced, meaning that a technical vehicle makes decisions on the basis of and under consideration of certain parameters. The use of artificial intelligence is only efficient if the decision-making process and the decision itself are no longer questioned by humans. If a program decides, however, whether a contract is to be entered into or whether a certain processing rate has been reached, the relevant employees may not even be able to achieve the agreed goals by their personal performance. Considering the current state of jurisdiction, this is unsatisfactory in so far as employees who are prevented from achieving their objectives by employer actions are entitled to 100% of the agreed variable remuneration, even if the agreed objectives per se have not been achieved. Where bonus systems also reflect the requirements of Section 315 German Civil Code due to the Federal Labor Court’s strict jurisdiction, the question arises as to whether a factual performance assessment by a technical system is able to represent an equitable discretionary decision at all. Of course, companies are first required to adapt their variable remuneration systems to the new technical circumstances and, where necessary, to modify their remuneration models; this example shows, however, that existing approaches to solutions are unsatisfactory.
Another potential problem for companies appears in a fundamental rights cover. As absurd as it may seem at first glance, there are areas in which employees are receiving professional instructions from artificial intelligence, so that a certain step in the process has to be followed by another mandatory step. As an example, it is undisputed under employment law that the performance of certain activities may be legitimately rejected on the basis of religious or ideological convictions. From this point of view, according to jurisdiction, devout Muslim employees may refuse to handle alcoholic beverages. The notion of rejecting specific instructions from human-like beings (artificial intelligence) is not far off then, when employees want to execute instructions only from humans as “God’s creatures” for reasons of faith. This aspect becomes highly interesting specifically because this conflict takes place in a constitutional environment. Companies will be held to organize their processes accordingly until this has been adequately regulated.
The use and development of artificial intelligence offers numerous operational advantages. Unless this development is accompanied by new legal guidelines, however, companies need to analyze their processes at an early stage to recognize the legal implications and implement appropriate instructions for action. Especially with respect to variable remuneration components, there may be a need for action, which is increasingly recognized by courts as well.