AI APPLICATIONS AND THEIR REGULATORY CONTEXT

The term artificial intelligence (AI) was coined in 1956 at the Dartmouth College Conference (USA) and was then considered bordering on science fiction. But especially in the early years of the new millennium, it has re-emerged in connection with increasingly evolved software. Today, AI can be defined as any learning technique of an operating system, which confers information not based on the human mind. It involves, in effect, a simulation of human intelligence processes by computer systems, whether it concerns learning, reasoning or information processing. Its applications are innumerable. AI already appears in all new platforms, applications or devices. It is also used in the interconnection of terminals and things in the IoT (Internet of Things). In particular, its use in the e-administration is increasingly relevant: in the areas of service management, education or healthcare, facilitating automated and algorithm-based decision making (e.g. the awarding of grants or taxation). It is permeable to any economic area and is incorporated into a diverse range of applications, which makes its spectrum very plural through machine learning algorithms. And among them, it contributes substantially to the development of robotics, of which AI constitutes, graphically, his brain.

Indeed, a unique application of AI is robotics. Robotics is concerned with the design, manufacture and applications of robots and combines various disciplines such as mechanics, electronics, engineering and applied artificial intelligence. It aims to establish non-human skills through integrated digital content. The robotics industry is constantly growing. Of the improvement of robotics, the following is a clear example Sophia android robot with Saudi Arabian citizenship.

The widespread use of robots, together with the significant degree of autonomy achieved, has generated a legal and ethical debate on the possible rules governing the responsibility assumed in their use. That is to say, who is responsible for the actions of a robot, and to what extent smart devices must take ultimate responsibility for their behaviour. At European level, as is well known, the Parliament adopted in early 2017 a report and recommendation aimed at achieving a higher level of legal certainty in the integration of robotics in the lives of European citizens. In addition, the Parliament, aware of the limitation of laws in this area, proposed a voluntary code of ethical conduct in the field of robotics for researchers, designers and users to ensure that the design and use of robots respects human dignity for the benefit of businesses and consumers. The next step was a public consultation, specifically targeting the future of robotics and artificial intelligence, with an emphasis on civil law standards. The publication of a new report should be available before the end of 2018. The difficulty of the subjects addressed, which have an impact on national laws, means that the report will be difficult to implement. Future EU legislative action will focus on three areas: the common European definition of an intelligent robot and the elements it should include; a registration system for advanced robots at EU level and what this should look like; and finally an EU framework for technological development based on social and ethical awareness. Leaving aside the ethical and philosophical debates on robot personality, - should the robot's personality be a part of a common European robot personality?electronic persons... - a splitting of the concept is possibly not unreasonable legal personality at various stages: natural person, legal person, robotics, just as there may be a certain personalisation of animals. However, this splitting up can hardly be dealt with in the framework of the Treaties, apart from intellectual property. There is a debate as to whether robots can be owners of intellectual property. At the present stage, the civil rules under which robotics could be subsumed are based on two fundamental areas: non-contractual liability and the protection of personal data, in turn linked to European instruments, which are possibly insufficient. Firstly, non-contractual liability is limited to the area of conflict, which is the subject of Regulation (EC) No 864/2007 (Rome II) on the law applicable to contractual obligations. This Regulation solves a problem of universal choice of law on the basis of the damage the autonomous concept of which shall be all the consequences resulting from a harmful act, unjust enrichment, business management or fault in contrahendo. The same insufficiency can be observed from the industrial perspective, where the updating of Directive 85/374/EEC is underway, thirty-three years after its approval, opening a debate on the concept of product, the liability it generates and the universality of European legislation. Finally, technological globalisation requires concerted action in international conventions, such as liability in autonomous vehicles, the clear application of artificial intelligence content in vehicles. As far as road accidents are concerned, there are currently no special provisions for autonomous vehicles. Regulation (EC) No 864/2007, mentioned above, is applicable, as is the Hague Convention of 4 May 1971 on the law applicable to road traffic accidents. The importance of the insurance sector in the possible satisfaction of the damage of the autonomous vehicle, and thus the content of the insured risks, is noteworthy. The IA is also essential in the development of drones with complex technological aspects. Their use may be subject to a special cybersecurity and aeronautical regime. Regulation (EU) 2018/1139 on common rules in the field of civil aviation establishes guidelines and limits for unmanned aircraft, whether or not they are considered drones. In particular, it has a bearing on safety and compliance with fundamental rights by their holders. Finally, we must be aware that the IA is still weak or primary. It is based on the current use of big data. When the technological leap is made to the creation of emotions or common sense, There will no longer be any reference to regulatory loopholes or to the neutrality of the legislator vis-à-vis technology. We should by then have clear universal ethical principles.