
Over the past few weeks, we have been looking at artificial intelligence (AI) in relation to data protection, copyright, trademark, and patent law. This week, we turn our attention to competition law.
AI can bring many advantages to a company. In the long term, it can save time, reduce costs, and allow employees to focus on the essential tasks within the company. This can give companies that use AI a competitive advantage. However, this advantage can only be exploited in the market if it complies with applicable competition law. But what does that actually mean?
What Is Competition Law?
Under normal circumstances, natural persons do not come into contact with competition law, and for many it is an unfamiliar and often overlooked area of law. However, it is competition law that protects the general public from unfair business practices, such as misleading advertising, false information on goods, baseless insults against competitors, and annoying telemarketing.
The following examples illustrate how AI can be used in the context of competition law and which approaches should be avoided.
Don't Hide behind a Mask
There is currently no general labeling requirement for AI applications. However, this may arise in the future from competition law.
AI-supported NPCs (“non-player characters”) have long been used in video games, and with technological advances, it will soon be impossible to distinguish between generative AI bots and humans. However, this is problematic when it comes to the comparability of online games. The use of many generative AI bots can make an online game appear more interesting than it actually is, as it creates the impression that it is a real competition with other users and that real contacts can be made.
If users are not informed that they are not dealing with real players, but only with very well-developed AI bots, this can have negative consequences for game developers who refrain from using generative AI bots or who provide proper information about them. Under these circumstances, this may not be compatible with competition law requirements. For this reason, it is advisable to indicate the use of AI to users and not to generate a competitive advantage from it.
Targeted Obstruction of Competitors
Even AI must comply with the usual rules of competition law and may not interfere with competitors outside of “normal” activities. This can happen through surreptitious means. For example, on social media, the success of profiles and posts is determined by the number of “likes,” “shares,” “follows,” or “subscriptions.” Business models involving purchased likes are not generated manually, but with the help of AI. This advertising with purchased or fake likes/followers constitutes unfair behavior.
On the other hand, it is also possible to hinder competitors through operational disruptions. This would be the case, for example, if AI were used to deliberately overload a competitor’s websites and servers.
Training is key
AI also violates competition law if it breaches legal regulations. This can happen during AI training if incorrect data is entered or information is omitted. This can then lead to incorrect results and wrong decisions by the AI, based on the principle of “garbage in, garbage out.” Amazon discovered this early on when they used AI to review job applications. Based on successfully hired Amazon employees, the AI recognized that the tech company’s workforce was predominantly male and identified a pattern. As a result, the AI rated women’s applications lower than those of their male competitors. Amazon was able to prove this error and tried to counteract it, but discontinued the project after three years of trials.
In summary, it should be noted that AI can intentionally or unintentionally violate competition law. Therefore, AI must continue to be monitored, just like an employee, to ensure that its actions comply with current legal requirements. To this end, it is advisable to establish effective AI compliance.
Next week, we will continue with the next blog post on civil law.
Datenschutzteam der BREDEX

Gastautor:in
Unsere Kolleg:innen aus verschiedenen Abeiteilungen veröffentlichen Beiträge als Gastautor:innen.
Jobs

Ihr:e Ansprechpartner:in
Gerne erzählen wir Ihnen mehr zu diesem Thema.



