Can AI be subjected to man-made law?

By Paramjeet Berwal

“Or will we learn that can they cannot be controlled at all?”

– Law Blogs Maastricht

If you have been following AI closely, you must have noted that efforts with regard to controlling, rather shaping AI landscape are directed towards ethics and not law-making as such. At least, this is the case, as of now. There is a difference between what ethics and law can achieve in contouring the AI-future and an article by Paul Nemitz, Principal Advisor in the Directorate-General for Justice and Consumers of the European Commission, clearly highlights the need for law making in the field. The Guardian also paints a similar picture highlighting that ethics are not enough and laws are needed to regulate AI. However, instead of speculating the future, I would like to go back to the very basic level and try to explore whether AI can be subjected to what we call ‘law’. OBLB’s post Robot’s Legal Personality, authored by Prof. Horst Eidenmuller, guides one in the right direction.

The question posed above reflects my hesitation in connecting ‘law’ to ‘AI’ because of, mainly, among other reasons, the fact that the very question not only defines but also limits what AI is considered to be, in futuristic sense. Controlling what is sought to be designed to achieve independent thinking, decision making and implementation is not only counterintuitive but also counterproductive. Though ‘human control’ is being advocated over the AI by some scholars, it is not likely to be the sure case, reports Financial Times. The ‘human control’ is what AI, inherently, seeks to go beyond from and have its own independent ‘intelligent’ existence. In other words, AI is supposed to evolve reaching not only human-level intelligence but also exceeding it. Scientists like Stephen Hawking of Cambridge, Stuart Russell of Univ. California, Berkeley, Max Tegmark and Nobel laureate Frank Wilczek of MIT, in their piece in Independent, boldly remarked that there are absolutely no fundamental limitations on what could be achieved in the context of AI.

Now, in the aforementioned context, let us analyse what law is. Law is regulatory framework that is made by humans in relation to everything that they control or that they think that they control. In fact, human beings have made laws for anything they can, have been able or seem or would like to control. We have made laws to control outer space, animals, water bodies et al but, in essence, the subject matter of that control is ultimately human beings themselves. Law has, at its core, basically two entities as its subject matter viz. natural persons and legal persons. Natural persons are human beings (people made of flesh and blood) and legal persons are companies and the like. The both have rights and duties but it must not be forgotten that concept of ‘legal person’ was created as a matter of convenience, for the humans needed to embark upon  entrepreneurial journey and desired some sort of free hand in order to take risks and manage liabilities. Companies were never perceived to have anything in common with human beings except for legal rights and responsibilities, in an economic context. In fact, Jefffery D. Clements, in his books Corporations Are Not People, argues very strongly against this legal notion of corporations being treated as legal person. Law is made by humans, subjecting themselves or corporations they are effectively in control of to it.

Let us look at the following in order to understand the scope of law, in its very essence:

Is there any law controlling the conduct of animals? No. They are not natural person; but, can they be “legal person”? No. Why? In order to be a legal person one must be subject to rights and duties, this 1928 Yale Law Journal article clearly demonstrates. Noam Chomsky in a talk at London School of Economics argued that animals could not have same rights as humans because lacking reasoning they cannot be considered to have responsibilities. He further elaborated his position, here. Therefore, animals cannot be legal person having same rights as human beings. In public discourse, one often gets to hear that animals should have rights. Steven Wise, Professor of animal rights law at Harvard and founder of Nonhuman Rights Project, has been advocating legal rights for animals. However, Richard A. Epstein, Professor at The Univ. of Chicago and Senior Fellow at The Hoover Institute, in his famous working paper ‘Animals as Objects, or Subjects, of Rights’, insists that animals should be afforded protection but the social institutions should not be radically transformed to recognise “the basic libertarian rights of freedom from human domination and exploitation.” It is pertinent to mention here that companies are considered to be legal person because their management are human beings that is responsible for exercising rights and discharging responsibilities that any legal person is subjected to. In other words, a legal person is so because it is under control of a natural person.

The next question that arises in this context is whether law can be made for AI or can AI be subjected, directly or indirectly, to law without totally altering what law really is to us? It depends on what AI is. What is AI? It is something else. It is, in the future, not only human-like but, possibly, more than human, in the context of intelligence or what it can achieve in unlimited dimensions. However, is it exactly human? Possibly not. If it is, even hypothetically, human, of course, it will participate in law making like us and be subject to it. But it is not human except may be in a very limited context of ‘intelligence’ and that too in a limited temporal dimension i.e. for a transient period in the near future. But this limited resemblance still doesn’t make AI human. However, then, can AI be compared to a ‘legal person’? Not at all, because it is capable of having a ‘thought-process’ of its own whereas companies are merely a ‘legal notion’ with absolutely no ‘intelligence‘ of their own.

In view of the aforementioned, AI, given the level it is expected to reach, is, in principle, very unlikely to be subjected to laws made by humans unless AI makes laws in order to govern itself. But this can only happen if the AI will never be able to go beyond the learning with regard to law making it acquires from human beings. Again, this speculation is very limiting in the sense that the very purpose of AI is to supersede human beings. We, the humans, do not know what the future will look like but the contemporary mind-set should entertain the ideas that only we, the humans, can do pending the exponential advancement in the field of AI.

Disclaimer: The views expressed in this post are that of the Author and do not necessarily reflect the views of the institutions he represents, the Tbilisi Tomorrow Institute or the organisations supporting the Tbilisi Tomorrow Institute.

Paramjeet Singh Berwal

Paramjeet, a lawyer, an invited lecturer, and AI research and policy consultant, is the Director of the Institute for Artificial Intelligence and Technology Law at Sulkhan-Saba Orbeliani University, Georgia. He is a Global Panelist with MIT Technology Review. He is frequently invited to deliver talks and presentations on various topics pertaining to AI. His research includes how AI will influence human existence, especially in the context of economy, work, law, society and its institutions, business management, social behaviour and policy making. He may be contacted at