LEGAL PERSONALITY FOR AI IN EU LAW
Amarbat Batzul
After Saudi Arabia ascribed the legal personhood and citizenship to AI robot called Sophia, it came into the mind of many lawyers, scholars and even practitioners. Since the AI began to create art including painting and music, we have to consider the possibility to give legal personhood to AI in order to protect its right and outcomes. However, this issue can be argued by many perspectives not only in EU but also in the world.
In EU regulations, the question whether ascribing legal personality for AI has been debated for many years in the higher stage of EU. In a European parliament resolution of 2017 with recommendations to the Commission on Civil law Rules on Robotics, the European Parliament urged the European Commission to propose what it called “electronic personality” for autonomous robots. However, number of experts have published an open letter[1] in April 2018 for Commission to ignore the Parliament’s move and reject “electronic personality”. The European Commission has published its press release on the future strategy on AI[2]. In this press release, European Commission only mentioned the liability, data protection and transparency issues that may be caused by AI technologies.
Therefore, In EU field, legal personality has not yet ascribed to AI technologies. Technically, EU has no power and authority to ascribe legal personality to anything within the Member States because granting the legal personality is the issue of national law.
However, I would assess the possibility to ascribe legal personality to AI technologies by two ways which are personhood and corporate legal personality.
In order to identify the issue of ascribing legal personality to AI technologies, we have to focus on the definition of what is actually AI. As defined in Britannica, AI is the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings[3]. The term “intelligent being” makes sense in terms of having a legal personality for AI. AI can learn something and repeat the programmed acts by themselves and develop such functions regardless of its inventors or creators. Thus, it is reasonable to debate over the issue of ascribing legal personality for AI. If AI has a legal personality, AI technologies can bear a legal responsibility and receives some obligations in terms of legal personality under the national laws. There are several models for granting legal personality for AI technologies.
First off, the question whether AI can have legal personhood as an “individual” is crucial to discuss. In other words, AI can have a citizenship like Sophia and considered as a “person” which will lead to the granting of human rights. Then we have to focus on the issue that whether it is possible in accordance with EU law. Ascribing natural person to AI technology may contradict with philosophical perspectives and EU laws.
From the philosophical perspective, only humans have dignity, the right to its integrity and the right to live. However, in this essay, I would not slide deep into the philosophical perspective.
In EU legal frameworks, according to Charter of Fundamental Rights of European Union and the Convention for the Protection Human Rights and Fundamental Freedom, the AI technologies cannot have a status of natural person because of the reason that only the human can clearly enjoy any rights. In my perspective, these charter and conventions are granting a right and declaring them to the humans, but not to other things. When it comes to the natural persons, the idea of consciousness should be arose in terms of legal personhood for AI. Human dignity and right to have integrity are the main concepts which include the concept of human is irreplaceable and every human being is valuable under the law. But for the AI technologies, in my humble mind, they can be replaceable and can be invented at any time. Based on this argument, I would say that AI technologies can never have a dignity and integrity because they are not biological humans. Also, an objection against granting legal personhood could be that robots lack any sort of feelings.[4]
Secondly, the legal entity theory can relate to this issue as well. There are many theories on corporate legal personhood, such as the legal fiction theory, the concession theory, the group theory, the real entity theory, and the nexus of contracts theory[5].
In the legal entity, it has some different perspective to consider the legal personality of AI. For example, there are always people who are deciding everything on behalf of the company. Behind AI technologies, there are also many people who develop and create such technology. However, the difference is a huge in terms of comparison with AI and legal entities. For the legal entity, although the liability is separate than its owners and shareholders, the owners or shareholders make the decision of that legal entity on how to act and what to do. However, the model of AI technology itself is different than legal entities. I agree on similar things, but the most important thing is that AI can act differently in regardless of impact of its inventors or developers. For this reason, AI technologies cannot be derived from legal entity theory when ascribing its legal personality.
The risk of ascribing a legal personality for AI is reasonable. As described by a lawyer Ugo Pagallo, there could be “double loophole argument” which means an ascription of legal personhood to AI could lead to two risks: humans could use AI to avoid liability, and AI could abuse human rights[6].
In conclusion, I support the idea of not ascribing legal personality for AI based on the argument by a lawyer Pagallo and experts’ open letter. If AI has a legal personality, many people or legal entity can assign their responsibilities to AI technologies in order to avoid any liability. Even it is unreasonable that how to punish AI if it does something unlawful.
In EU stage, everything will depend on national laws of Member States. On the other side, the complicated legal relationship between Member States and European Union definitely affects the possibility to ascribe legal personality for AI. More specifically, EU Parliament or Commission do not have any authority to declare that AI has legal personality, also, the Member States cannot grant AI a citizenship or legal personality in terms of single market rule of EU.
[1] Available at www.robotics-openletter.eu
[2] “Artificial Intelligence: Commission outlines a European approach to boost investment and set ethical guidelines”, European Commission-Press release, 2018, available at https://ec.europa.eu/commission/presscorner/detail/en/IP_18_3362 (last visited 20.07.2020)
[3] “Artificial Intelligence”, B.J.Copeland, Encyclopaedea Britannica, 2020, available at https://www.britannica.com/technology/artificial-intelligence
[4] “Robotics, AI and the Future of Law”, M.Corrales, M.Fenwick, N.Forgo, Springer, 2018
[5] Migle Laukyte. 2019. AI as a Legal Person. In Proceedings of International Conference on Artificial Intelligence and Law (ICAIL 2019). ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3322640.3326701
[6] Migle Laukyte. 2019. AI as a Legal Person. In Proceedings of International Conference on Artificial Intelligence and Law (ICAIL 2019). ACM, New York, NY, USA, 5 pages. https://doi.org/10.1145/3322640.3326701