CHAPTER 1
Introduction to
Artificial Intelligence
and International Law
The advent of AI is not a nightmare… It is a differing category based on a newer league that manifests its creation. It is considered to be a generalization from a bigger leap of ‘knowledge machine’ to a more mature explainable entity, which is a great marvel to dream of. This is not an easy journey either. It takes a good time to make up and attain the status quo that we have reached hereby. History gives us a befitting chance to look upon the historical aspect of AI for a due developmental purpose. However, such a development, where content and identity become gross values of impeccable relativity for human dimensionality, it is obviously relatable why Stuart Russell believes, “Humans are defenseless in information environments that are grossly corrupted.”(itut, 2017). However, the story of AI for a law de lege ferenda (a new law) for an information-oriented society is not as direct and simple as it seems. In addition, there is a time when AI needs to reconfigure and properly understand the law and so does the law too. So, is it just the aura and questioning concerns of the common law, which is obviously considered as one of the most flexible innovative ways of the development of legal development? No. Common Law, a part of legal instruments and literature (originated primarily as a part of English Law), which is based on precedents and customs, is a league of those tools that instrument and relearn from the society and also provide insights for the society to learn. However, if we have a proportional understanding of these, we will come to know that this manifestation has a tilted problem, which arises when the law fails to entail self-transformation. Common Law needs more learning rather than the society to recognize and relearn things. Thus, there is a need to find the neutralized, less-biased or bias-free common legal approach, which common law furthers. This is understandable from the approach of Lord Denning in a case where he propounded a modular structure of international law in the generic sense of common law doctrine; however, being beyond the limits of the legal aspect so forth. He endeavored to modify the absolute principle of ‘sovereign immunity’ into a preventive principle whereby it is understood how some instrumental generalizations can be brought up and legal sovereignty in its intrinsic and extrinsic senses are beautifully and responsibly demarcated (Rahimtoola v. The Nizam of Hyderabad, [1958] 1 A.C. 379, 1958). Denning’s approach to international law was somehow coined by scholars of international law in the ambit of monism and dualism. While monism focuses on conjoining a national law of any state with the international legal framework, dualism separates them apart. This becomes important because these legal tenets became the original basis for nation-states to democratize the human society in their own way, paving ways for globalization and advancing technology, later to a bigger asset for the welfare of mankind. In addition, the story does not end. It progresses and breaks the barriers of legal positivist1 thoughts, which equate the legal value as supreme in a questionable sense. That is where a bridge of correlativity is formed between international law and the laws of various nations (especially those having common law). So what sense does it provide for the motivation of a differing field called Artificial Intelligence, which is based on machine reception and activity basics and their intrinsic aspects?
“We’re going from a world where people give machines rules to a world where people give machines problems and the machines learn how to solve them on their own (Goldman Sachs, 2016).”
AI is not a field that law cannot pursue. In addition, its relationship and development does not limit to data protection legislations, regulations, orders and other legal instruments. It is connoted with human development and activity – a special part of the human rights doctrine but also a special part of the welfare and societal privacy policy. Perhaps that is the reason why Facebook has been condemned for its policy loopholes and business models with respect to data and AI development. One of the most instrumental voices that emerge to condemn and expose the AI business planning is a TED Talk by Zeynep Tufekci, an ardent critic-techno-sociologist (TED, 2017). That is why if we understand an opinion given by Lord McNair in the West Africa case, then we again get the same reflection for pursuance.
[The] way in which international law borrows from this source is not by means of importing private law institutions ‘lock, stock and barrel’ readymade and fully equipped with a set of rules. It would be difficult to reconcile such a process with the application ‘the general principles of law.’. (Status of South-West Africa case, ICJ Rep. 1950 148, 1950)
There is a need to form bridges to bring international law out of the dependence of the human rights approach to provide a newer and wider scope of legal innovation to recognize the roles and implications of Artificial Intelligence. However, the road is not planned and nothing is set up, except the material cyber obligations and regulations that provide an extrinsic insight.
So, if we go back to the realm of Artificial Intelligence, there is a grip of understandability that makes up the insight of data reception. And it is also about how the machine works and makes things more prone to reception and reaction and the capillaries – the learning part – a whole new league of understandability. Now, the role of artificial intelligence is more entitative. It is not just about the design that it has, the algorithmic structure it is based on, or the data that it has to process. Predictability and data, the external manifestations of an AI, in an information economy have become wider and have the potential to change the way in which our human artifacts exist in a wider sense today. The whole summation or product that comes into understanding is the purview of AI Ethics, the field which regards the all-comprehensive potential of artificial intelligence in various sectors such as finance, healthcare, education, law and others. Also, this concern produces the insights and questions related to the material aspects of AI such as the responsibility of the companies and researchers, their business models, data sovereignty and the data regulation mechanisms that make an exclusive form of learning possible for people. This book covers the materials and immaterial aspects of Artificial Intelligence with International Law and furthers its horizons in different aspects.
ARTIFICIAL INTELLIGENCE: A DILEMMA FOR LAW
Is the development and activity of AI so difficult that the Law cannot understand it? Well, that is the premise for questioning the due concerns so as to understand how it progresses. It is not a simple model and it is obviously a dilemma because certain modalities need to be clarified. Let us take a dive into the philosophical constructions that manifest and create it in generic standards. When the idea of ethics was under consideration to monitor and understand how the human society works, this was a clear premise that ethics was the technical instrument to be developed. Aristotle assumed without an argument that we acknowledge rather very general but these typical parts are such that they represent a symbolic resemblance and value in the case of a reasonable discourse and is considered far more rational in some terms (Moss, 2015). It was merely a binary of categories such as a league of syllogisms accrued with human Diaspora with certain limitations as the practical dimensionality of reality. The two-dimensional conceptuality has grown into a three-dimensional idea where rationality found a spectator in a prominent role. The eminent Turing Test is all about this. Its pre-context, as succeeding to be asserted by Descartes, is a pre-furtherance to understand the ethical and practical dichotomy that still exists when it comes to understand AI for Law:
[H]ow many different automata or moving machines can be made by the industry of man [...] For we can easily understand a machine’s being constituted so that it can utter words and even emit some responses to action on it of a corporeal kind, which brings about a change in its organs; for instance, if touched in a particular part, it may ask what we wish to say to it; if in another part, it may exclaim that it is being hurt, and so on. But it never happens that it arranges its speech in various ways in order to reply appropriately to everything that may be said in its presence, as even the lowest type of man can do (Descartes, 1996, pp. 34-5).
So, if there is a simple idea which says that Artificial Intelligence is not a fake reality but a manifestation to re-initiate visualization of reality and its concerns, it is understandable that if we take one side of human observation from the interface of AI, then it is very clear that we need to have some dimensions taken into a legal aspect on a theoretical and practical basis. This is certainly not a simple piece of cake as it seems and some abilities can be genuinely ascertained. There are two basic dilemmas attributed to the approach of AI for legal recognition: (a) The human rights doctrine and (b) The privacy doctrine. Both of the doctrines have a due concerned matter of establishment and process that defines human society and replenishes its dynamic personal and interpersonal history. This is the basic outlook that AI systems must understand if it is a deemed figurine that a human rights approach can help AI to develop and reinvent. The Privacy Approach, however, is clearer and anticipating because it does not confine AI to the two-dimensional right-duty or obligation-observation approach but increases its scope to the realm of a hidden receptivity, which maintains the paradigm shift of human resourcefulness and pragmatism towards affording potential solutions whether in business, science, administrative or legal affairs. This question is addressed in the following chapters.
Human rights, however, is a binary concept where two or more entities are treated under relatively bi-polar recognition. In classical civil legal concepts of human rights, Hobbes, Locke, Rousseau and many other thinkers compared the state-public dualism with the urban-nature realms and put forward their own rationality-based or preconceived ideas with respect to the state structure and the civil rights. Civil society in older times was just a linear image. This linear image had little scope of dimensionality as legal and civil thought was still based on cause-effect relationship. It conceived a two-dimensional image when we entered into the age of contemporary international law. It was not Grotius, but Kelsen then.
Recognition of international law generally as a valid body of rubrics has instrumented itself as a steady procedure (Oppenheim, 1992, p. 3). It is nevertheless distinct that States have developed more and more conditional on each other, a kind of singularity maybe mainly attributable to the rising ‘institutionalization’ of the global community (Sztucki, 1974, pp. 35, 165). Such conditions decide the way nations endorse trade among themselves, make up their diplomatic and consular relations, and play their part in various regional or global alliances, whether military, economic, humanitarian or of any other type. So, what is the natural context to further upon when it comes to dealing with the discourse of human rights?
Human rights in general are not the final discourses of international law; but for artificial intelligence, they are quite instrumental and important and they manifest certain exclusive aspects, which are immaterial and instrumental. AI development has a receptive furtherance to human society as it is in organized sectors. In the technical aspect, it may seem differential but this is not the ultimatum behind the prerogative, and so where AI realms on Symantec, MakeMyTrip, Alibaba, Tencent, Google, Facebook, Twitter and other business realms are used in the sense of data specificity, quality and other legal and illegal dynamics, this may represent a material element of human rights, of which one of the most high-profile cases is the antitrust fine imposed on Google (Chee, 2018) by the European Union. This represents a linear aspect of legal imploration of data interactions and activities, which matters in the core context. In the bare view, it seems to be the material aspect of international law (rather a private one for matter of representation). However, the immaterial aspect of data, AI or any such cyber realm is not touched, which is only entitled by ISPs, GPS sensors, OS tech and other components. Even Blockchain has its own categories, where the technical aspects are dimensionally neutral but the applicative horizons are not similar or even so same.
Thus, it is a deemed necessity to understand the dystopia that AI brings up when it forms or provides extra dimensions to human rights. That is how the league of linear human rights end and privacy arises from a basic aspect of being an inherent and rather philosophical human right comes into play, becomes more shaped and integrally important, dimensional, real and critically sensible. Let us understand this aspect as well before moving to the aspect of AI.
The development of CSR, in comparison to the polluter pays principle, is under the same prerogative of the privatization of law, where the corporate liability is converted into responsibility. Polluter pay is a private law, which focuses on imposing serious liabilities on industries to comply with environmental preservation and protection, but CSR has made minimal financial criterion for companies to bear ethical obligations to commit towards social welfare. And let us be honest. Liability is a material character in the classical sense and is derived and furthered by sanctions, while responsibility is inherent and true. It is reflected, understood, realized, born and dies at its own upheld prerogatives entitled with it. That is why the concepts of strict liability are just sanctionable and do not resemble always. Till today, liabilities exist as crude sanctions, but not all sanctionable liabilities may be useful. What law should do to benefit people is that make civil obligations reflective of social values, rather than just with a purpose to showcase justice and punish people. However, the crudeness of liability is not the gift of jurisprudence; it is due to obscurities in social life based on ages. Thus, it is a necessity as was in the case of contemporary international law to develop out of simple aspects of liabilities, consents and classical principles towards more open, diverse and accepting principles for newer developments at public and private level to benefit human development for a better future. Nowadays, such minimal obligations are used in the international human rights treaties to make sanctions and rights purposive rather than retributive.
Thus, this is the just the beginning of the bettering aspect that manifests the due realm. The next sections of the chapter shall deal with the related perspectives.
LEGAL LINGUISTICS: A PATHWAY TO MODERN LEGAL CONCEPTUALIZATION OF AI
Law is befitting to be benefited by ethics and linguistics. In Artificial Intelligence, there have been frequent ways by which this generic aspect has been successfully covered. This becomes instrumental and important in a simple context because it manifests the basic blocks of law and legal reach for any subjective or objective development in the human society. In a sui generis or generic sense, legal linguistics provides a back-end to redefine the structural and inherent attributions of Artificial Intelligence and this becomes an important end to the purpose (Ashley, 2017). Language works in an awesome way enough to let artificial intelligence open perspectives in its method of working and producing. Machine learning has been thus innovated for a genuine utility and has manifested an ad hoc manifestation for human and other material activities. It covers an extensive realm of data receptivity and understandability. However, its opaque nature deprives itself from activity and interaction and the furtherance of data receptivity, which is essentially important for the concern of review. If AI via machine learning fails to self-transformation and is caged in the numeric possibilities that make it vulnerable and highly susceptible to prediction, then it may lack values in terms of what it can do to be an essential part of human society, in whatever form it is.
It is claimed that machine learning (ML) and deep learning (DL) (Hof, 2013) estimate human intelligence. However, presently these tools essentially spot designs that are meaningfully modified by humans and necessarily be construed by human instruments to be beneficial. Concluding, the developments duly represented are evolutionary and not radical. One grave restraint of ML is that as a data-determined method, it essentially relies on the value of the causal data and thus can be very inelastic (Cummings, Roff, Cukier, Parakilas, & Bryce, 2018, p. 13). This sometimes becomes entirely important as if this consideration is really not adequate. We may be subjected to some consequential aspects of AI, which might not be correct enough for a matter of receptivity and there shall exist such barriers between humans and them. Hence, there are so vague settlements so as to the receptivity itself.
But there is a better note for view. The basic aspects of legal linguistics to understand AI and the AI in case of legal linguistics are exactly the same, where there is a need of a dimensional reciprocity at least. Let us experiment this aspect. Suppose that there is a statute of some nation (say S) having one of its law in the form of a statute namely the X Act. Now, if we suppose that an AI System (Say I) has to learn the interpretation, mechanism and jurisprudential extent of the Act, then I have to learn certain important attributes that make its basic building blocks. They can be possibly:
| 1. | Scope, extent and jurisdiction of X |
| 2. | Amendments, Case Laws or Precedents related to it (directly/ indirectly, based on scope of Territorial Division and other ... |
