دانلود کتاب هوش مصنوعی و مسئولیت ظالمانه

Hardly can any discussion about the future occur without reference to digital technologies. Whether seen as a utopian vision or a dystopian concern, nearly everyone has thoughts on the impact of digital technologies on our lives. Among digital technologies, artificial intelligence (hereinafter: AI) is the one that captures the most attention and raises eyebrows. Often described as unprecedented and groundbreaking, it bears the potential to change the way we interact with one another, do every day simple and complex tasks, and, eventually, how we perceive the world. AI would be able to incrementally replace humans in their jobs and probably so in an unmistakable fashion. If AI replaces human jobs, what would be the fate of the vast workforce left jobless? Can we already now predict who are going to be victims of the change and do something to enhance their position? Historian of technology Carl Benedikt Frey in his book “The technology trap”1 suggests that we have seen this before and that inequalities in wealth caused by the emergence of groundbreaking technology were a realistic scenario in the past. The rise of inequal-ity therefore appears as a recurring pattern in technological revolutions. Addressing the breadth of potential problems in this work would be overly ambitious. Rather, I will focus on the responses of the tortious liability system of Bosnia and Herzegovina (hereinafter: B&H) to accidents involving AI.
This, too, is a question which, in my opinion, deserves the attention of not only lawyers. The practical implications of having AI involved in potentially harmful events have broader societal importance. Reduction of human involvement in accidents caused by human errors, such as fatigue or stress requires rethinking legal standards with characteristics of humans in focus. Technology, unlike humans, doesn’t experience fatigue or stress. It does not have a sick child due to which it would lose focus or simply have a bad day. Research suggests that a great number of motor vehicle accidents occur due to human errors, such as speeding, driving under the influence of alcohol and other psychoactive substances, nonuse of motorcycle helmets, seatbelts, child restraints, and distractions such as mobile phones.2 None of these problems may exist if we exclude humans as the operators of vehicles. However, implementing AI in activities requires rethinking how tort law systems work. Challenges within the system exist on both broad and specific levels. On the broad level, for example, it is hard at the outset to formulate what is AI in an operable legal definition. There is no consensus among scientists on what is AI due to its changing and unpredictable nature. Definitions typically describe it through terms such as autonomy, intelligence, and opacity, but even these can be vague. What is autonomy or intelligence exactly? If an acceptable scholarly definition of this phenomenon may be hard to fathom, so is a legal one. Answers to many questions should not be a matter of exclusive deliberation of lawyers but of a broader population. For example, whether AI should be allowed in every field of our lives, or certain parts of it are simply too vulnerable or sensitive. This has been debated in the EU institutions as a part of the broader discussion about the legal instrument titled AI Act. While such questions do not strictly fit under the work on civil liability, they certainly cannot be divorced from it. They will be at least mentioned in this work, as well as their importance for the regulation of liability. Another set of questions that can be described as broad has to be addressed with a clear opinion of their author, however. Those are broad questions that act as a basis for any further deliberations on the regulation of liability for AI. In this work, such questions are provided with a clear understanding and proposed solutions. The examples would be understanding of product under the regulation of product liability and which of the potential addressees of liability deserves to bear the ultimate burden of remuneration.
Despite the title puts Bosnia and Herzegovina in focus, this work relies on the work of scholars from different countries. There are two reasons why this approach is not only justified but only possible. First, as far as I am aware, jurisdictions around the world are still developing their approaches to the regulation of AI liability. The non-existence of specific rules on AI liability does not mean that there are no rules on liability at all. A decision to treat AI with general rules of liability, such as the ones found in civil codes and statutes, is also a regulatory choice. General rules are designed to be future-proof and technology-neutral in order to be flexible and therefore applicable to a range of accidents. While general rules around Europe differ from one another, they are typically based on the same foundations and principles. These similarities allow for drawing comparisons. Second, scholarship in Bosnia and Herzegovina (hereinafter: B&H) on legal aspects of AI is almost non-existent. It would be impossible to use the opinions of local experts since they are unavailable in written form so far.