The new law of artificial intelligence
Prof. Dr. Janine Wendt, Prof. Dr. Domenik Wendt
Das neue Recht der künstlichen Intelligenz
Nomos Verlag
1. Auflage 2024
330 Seiten, 59,00
ISBN: 978-3-8487-8980-1
Some will ask themselves the question: How can I become familiar with the EU AI Act (EU 2024/1689) and gain expertise that I urgently need in my consulting practice because the demands from clients or my employer are increasing? The answer is not easy, because the subject matter is complex and requires a certain amount of familiarization with the legal text and the circumstances that have led to regulation. One thing is certain: there is no way to learn the subject matter in your sleep, but a good textbook can help you prepare.
The newly published textbook by Professor Dr. Janine Wendt from TU Darmstadt and Prof. Dr. Domenik Wendt from Frankfurt University of Applied Sciences is ideal for this purpose. In a very readable 160 pages, it provides a comprehensive introduction to the subject matter and touches on all relevant topics. The remaining 160 pages of the textbook are filled with the text of the regulation, which is very helpful and makes reading even more enjoyable because there is no need to consult several resources. The authors offer a clear and differentiated presentation of the subject, which is characterized by its conciseness and focus on the essentials. A particular advantage of the book is the critical and constructive evaluation and analysis of the AI Act. The authors correctly point out that we are currently still at the first development stage of ‘weak AI’. Most AI systems on the market therefore come with little or no relevant risk. This approach is refreshing, as it contrasts with the sometimes exaggerated and unreflected statements and debates about the risks of AI.
It is noteworthy that the authors criticize the horizontal approach of the AI Act, which is actually not without its problems, as the use of AI in different sectors is associated with very different levels of risk. Furthermore, the book is characterized by a sense of proportion and realism, which sets the two authors apart from other, sometimes perceptible voices that paint the AI devil on the wall. As the authors rightly point out, this does not exist because we are still a long way from the so-called superintelligence of AI systems, the first manifestations of which we can expect in 20 years at the earliest.
“Tatsächlich wird von der überwiegenden Anzahl der gegenwärtigen auf dem Markt befindlichen KI-Systeme nur ein geringes oder gar kein Risiko ausgehen.”
Against this backdrop, the textbook helps to avoid throwing the baby out with the bathwater without negating the risks. With reference to the EU Commission’s forecast, the authors emphasize that high-risk AI systems that have a potentially detrimental effect on people’s safety or fundamental rights only account for 5 to 15 percent of applications on the market.
Overall, “Das neue Recht der Künstlichen Intelligenz” is a highly recommended work for anyone who wants to get to grips with the topic of artificial intelligence. It offers a clear and differentiated presentation of the topic and is an important contribution to the discussion on the regulation of AI. The authors have contributed their expertise in this area and offer a well-founded analysis of the relevant laws and regulations.
