Yapay zekânın hukukî statüsü ve hukukî sorumluluğu
Özet
The issue of granting legal personality to artificial intelligence, in essence, refers to a decision to grant a set of rights and related obligations to that entity. There are some basic questions that should be answered by especially information technology law doctrine and practice, regarding which criteria should be sought in the process of establishing a legal policy for the recognition of non-human beings and transforming this legal policy into a normative regulation. The starting point in solving the problem of whether an entity can be recognized as a personality is to determine the meaning, scope and legal nature of the concept of personality. In the second stage, the entity, which is envisaged to be granted personality rights, is subjected to an evaluation process within the framework of the material approach, which sees the personality as an existential structure, and the formal approach, which is based on whether the law and society ascribe personality to an entity. There is no doubt that systems with a limited scope of activity and autonomy, defined as narrow or weak artificial intelligence, should be accepted as objects by the law, depending on these characteristics. On the other hand, the level of success reached by cognitive technology today has also allowed the development of autonomous artificial intelligence, which can learn from its own experiences through machine learning with different algorithmic structures and complex software and can act independently without any human interference. The autonomous decisions and actions taken by the artificial intelligence during the fulfillment of the tasks defined for it can sometimes damage the assets or personal assets of individuals or cause a breach of contract in obligatio. In this respect, today, the need to develop a unique personality model has emerged in terms of artificial intelligence beings with a strong autonomy feature. The legal liability realized in terms of systems that are expressed as narrow artificial intelligence and whose autonomous feature is very limited will be shaped within the framework of contract, tort, product liability or consumer protection law. The type and characteristics of liability will differ according to the use and place of artificial intelligence. According to positive law, this type of artificial intelligence is subject to the same provisions as other assets that are the subject of law, depending on whether they are considered as objects or products. In this respect, in the event of a damage caused by narrow artificial intelligence, the legal responsibility will not belong to this artificial intelligence, but to those whose actions the laws impose legal consequences due to these entities. The liability arising from the actions of artificial intelligence with strong autonomy is incompatible with the general principles of liability law due to the unique characteristics of these entities. Since some of the damages caused by these entities, which have a strong autonomy, cannot be compensated by the current provisions and they do not have their own legal responsibilities, a legal gap arises in terms of compensation for the aforementioned damages. For this reason, a unique type of strict liability should be established based on the algorithmic structure and autonomy feature that gives the new generation artificial intelligence its characteristic feature. This strict liability should not be based on one of the current strict liability principles, but on a specific responsibility basis under the title of strict liability arising from the "autonomous behavior of smart machines".