As a lawyer, or a person involved in the legal profession, the maximum interaction one has with technology is via MS Word. This single innovation changed our lives and we upgraded from typewriters to computers. Add PDF Readers and online legal databases and that completes a legal professionals technological arsenal. Typewriters, however, are not fully redundant, as some lawyers still opt for them.
With the onset of the fourth industrial revolution and the rampant growth of Artificial Intelligence, the world, as we knew it changed. The lawyers got the gift of tracking changes and sending emails and for the longest time, we were happy with that. But now, with further development of artificial intelligence, almost to the extent of near-human intelligence, a new idea has come to fruition, that is, legal ontology.
Before going into the concept of legal ontology, it is important to understand what the term Artificial Intelligence means in the legal field. When one hears the term Artificial Intelligence, the most common understanding of it is robots and self-aware machines. In the legal field, however, the meaning is not as interesting, rather it relates to cognitive computing. This means teaching a computer how to learn, reason and make decisions. This is done by focusing on the patterns that exist in the raw data and then testing that data to get the results. This would have huge implications on the corporate world, especially concerning analysing contracts and completing due diligence. This could also result in most paralegal and research assistants becoming obsolete.
What then, is ontology? It simply means representing knowledge in a machine-readable form. Legal ontology aims at converting legal text into a standardised machine-readable form that computers can comprehend and thereby analyse. Whether it is for contracts, legislations or any other legal information, the aim is to make it machine-readable to an extent that the Artificial Intelligence can give precise answers from a set of predetermined answers. In the legal field, take out the jargons, the principle remains constant, and it is that principle that the Artificial Intelligence needs to be taught, and then just tweaked from time to time.
This paper analyses this concept along with a brief description of how this has successfully been implemented in India, the Intellectual Property safeguards available to such Artificial Intelligence and the long-lasting implications of such technology in the field of law.
The concept of the term ontology is simple, as mentioned earlier: teach your knowledge to the machine. This could hold a lot of benefits, especially in the field of contract law and corporate law. To err is human, and legal compliances and due diligence often leave room for human error. This can easily be prevented by integrating Artificial Intelligence in legal practices. It can even find implementation in the field of civil and criminal law.
One of the most affected areas would be that of contract laws. This is mostly because most contracts have a standard format, which means that there are standard patterns that can be taught to the AI, thereby making the process of drafting agreements and cross-checking them fast and cost-effective. For instance, a standard Non-Disclosure Agreement will have certain fixed clauses such as a non-compete clause, definition of confidentiality, obligations of the parties, exclusions, dispute resolution clause, etc. Such clauses form a pattern that can be fed to the Artificial Intelligence and the Artificial Intelligence can then be taught to understand these clauses by feeding data from different types of NDAs.
Due diligence is a term every lawyer knows and every law intern dreads. The process of due diligence is again set in fixed laws with fixed patterns and hence can be taught to a machine. Legal compliances are very time consuming and leave a huge margin for error, hence with such technology, legal processes can become faster and the margin of error is greatly reduced, will reduce litigation.
In this field, Artificial Intelligence can be used to conduct interviews as it has been observed that people are more likely, to be honest with a machine rather than people since a machine is not capable of judgement . In this way, the process of litigation can be made faster and more cost-effective.
Before dwelling into the practical aspect of legal ontology it is pertinent to understand the jurisprudential aspect behind it. This is most clearly understood under Classical Legal Positivism. Kelsen’s Pure Theory of Law forms the basis of legal ontology. Although this theory does not directly speak of ontologies, it forms the crux of the jurisprudential part of the legal ontology. The most relevant jurisprudential basis behind the legal ontological theory is Alexy’s Theory of Fundamental Rights , which proposes the “weighing and balancing structure” which essentially involves two aspects: norms as rules and principles and legal positions.
The first aspect of the theory is about legal norms. Norms, in this case, are further classified as deontological norms and axiological norms. Axiological norms lay down principles relating to the value of a thing. Deontological norms are then further classified as rules and principles. Principles have different degrees of fulfilment depending on factual and legal aspects whereas rules are norms, which are either fulfilled or not.
The second aspect is the legal positions. Alexy himself has defined legal positions as situations in which a subject, in a legal relation, has a right against another subject. This concept has a complex structure where he has said that every legal position is a relation between two subjects and an object. Accordingly, he further divides rights into three parts: rights to something, liberties and competences. Rights to something is once again divided into rights to negative acts and rights to positive acts.
Rights to something, according to Alexy can be represented through the jural correlatives and jural opposites given in Hoffeld’s Theory. Jural opposites are: right-no right, Duty-no duty and jural correlatives are: right-duty, no right-no duty.
With a brief understanding of Alexy’s Theory, the relevancy of this theory in the field of legal ontology lies in providing a basis of the framework in establishing logical connections between legal propositions which can then be converted into code. Raw legal data could not be taught to a machine due to its subjective nature and that is when this theory found applicability in a legal ontology. These legal theories also helped classify ontologies.
In this regard, a foundational ontology defines a set of domain-independent categories. Core ontology, on the other hand, is far more applicable in the field of legal ontology since it defines a set of fundamental concepts of a field of knowledge, for example, law, organisations or software. Often core ontologies are built by extending a foundational ontology. Core ontologies that represent legal domain-independent concepts are called Legal Core Ontology (LCO) or Core Legal Ontology (CLO), which has been discussed at length in the next segment. Some of the Legal Core Ontologies will be discussed here. The first being Valante’s Functional Ontology of Law . According to this, the main purpose of the law is to control social behaviour.
The next Legal Core Ontology is developed by Kralingen and Visser . It is a framework based ontology of law which is then decomposed into the generic legal ontology and statute specific ontology. Generic legal ontology divides legal knowledge over norms, acts and concept descriptions. A statute specific ontology also further divides the generic ontology to derive the fine details of individual cases. Kralingen and Visser emphasise on the distinction between physical acts and institutional acts.
To understand the rampant growth of Artificial Intelligence, Gordon Moore, a scientist at intel formulated what is known as Moore’s Law. His law predicts that this trend will continue, and growth in computer power will double roughly every two years while the cost of that computing power will go down . The advantage that this possesses is that humans will not be tied down by tasks that a machine can do, thereby increasing overall productivity.
The first step to cognitive computing of legal data is Core Legal Ontology (CLO). CLO is essentially a constructive ontology since it since it allows to reason over the contextual constraints that can be intentionally adopted by a cognitive agent when recognizing or classifying a state of affairs .
CLO is aimed to provide three kinds of legal tasks: conformity checking, legal advice and norm comparison. Conformity checking relates to the representation of social and legal situations concerning legal norms. Legal advice is a more complex form of conformity checking wherein the concept of common sense comes to play. This makes it more difficult to take abstract concepts such as reasoning or interpretation issues and convert it to an ontology. Norm comparison relates to conflict handling, another aspect that is not simple for a machine to comprehend. These three aspects make up the essential structure of the legal ontology. The mode to bring this to fruition is developing programmes that then need to be taught, via examples. This is where engineers and lawyers come together, to create a machine capable of such understanding.
The most baffling question that arises is can reasoning be taught to a machine? It can be, with a lot of complex programmings such as Python or DOLCE. This, however, would be in the domain of engineering and beyond the scope of this paper.