New Technologies in International Law / Tymofeyeva, Crhák et al.

As much as the intentions of the developers might be good and the AI technologies should be only benefitting to mankind, there might be a risk of unauthorised change of programming of the software operating for example the artificial limbs or software which communicates with human brain. Therefore, there should be some sort of legal rule, that these AI technologies must have a security measure, which in case the software gets hacked, the artificial limb or medical device (for example a wheelchair or a lift operating based on AI) will immediately stop operating and the human service operator will be automatically called to check the device and fix the problem. It might be also discussed, whether there should be accepted a legally binding rule, that the artificial intelligence must not change the physical nature of a human being. In the author’s opinion, there might be strong ethical grounds for such prohibition. Conclusion In order to fulfil the aim of this paper, there are answered the two research questions. In connection with the first research question focused on examining on whether and what the human rights could be possibly violated by the AI systems, following can be stated. In the chapter three, there were mentioned the major areas of human rights, with which the AI systems might interfere with. Those were prohibition of discrimination, right to privacy and right to work. This includes the systems that are used in military, security as well as civil context. The potential risk of human rights violation is mainly due to the technological aspects of these systems and the options as well as outcomes which stem from their use. The author also noted in the chapter three several areas of human rights to which the AI systems might have benefit impacts due to the technological possibilities of these systems. Those are the right to an adequate standard of living, the right for the continuous improvement of living conditions and the right to achieve the highest attainable level of physical and mental health. The more these technologies are expanded and used in various areas, by public authorities for civil, security and military purposes, as well as by private individuals, the greater will be the need to ensure that a large-scale and serious violation of human rights is prevented from happening. This might occur due to deficiencies in the functioning of these technologies, especially if the decisions taken by the systems are not verified by responsible human operators. In connection with the second research question focused on examining on how these technologies should be used, so they do not interfere with the existing laws of human rights, following can be stated. The AI systems should always remain under human control mainly due to the prevention of creating situations which are not regulated by any currently valid laws. The existing legal rules stipulating the issues of accountability, responsibility and liability are not applicable to any technologies operating on the basis of the AI. There should be clearly stipulated the rules of responsibility for the human rights violation by using AI, even if the AI systems were not directly ordered by a responsible human commander or operator. There should also be established universal legal framework for the human rights protection in relation to the use of the AI. Among other things, there should be performed regular legal assessment of impact of AI to human rights. This legal framework should also include a control mechanism. One of the options for the control mechanism might be creating an organisation which will

208