Bridging Legal Divides: The Imperative Of India’s Legal Ontologies/knowledge Graphs

Introduction

AI-powered legal research has emerged as a frontrunner in the legal tech landscape, and many believe it represents the future of the field. However, the question remains: is the current technology robust enough to deliver consistently efficient and accurate results? A recent case involving two lawyers found guilty of citing bogus case laws generated by AI highlights the limitations of current systems. While these lawyers blamed ChatGPT for tricking them, the incident underscores the crucial need for AI to develop a nuanced understanding of legal complexities, context, and specific case law to generate accurate legal searches.

While current AI technology struggles with verification and can generate inaccurate information, there are solutions to overcome the peril. Ontologies and Knowledge Graphs are powerful tools that can help validate research results, ensuring a higher degree of accuracy and reliability .

In this blog post, we willexplore the intersection of AI-powered legal research and India’s intricate legal landscape. In the first part, we will first delve into the challenges faced in automated legal research within India, examining its complexities and the imperative for robust technological solutions. Then, in the second part, we’ll pivot to a discussion of the transformative potential offered by Legal Ontologies and Knowledge Graphs (LO/KGs) in addressing these challenges. Additionally, we’ll explore their applications within foreign jurisdictions and return to India to discuss their role in building a more inclusive legal infrastructure.

Part I

India’s legal landscape presents a unique challenge for automated legal research due to its inherent complexities. Navigating the vast and intricate web of legal information for efficient research can become difficult for several reasons.

Jon Brodkin, ‘Lawyers have real bad day in court after citing fake cases made up by ChatGPT’ (arsTechnica, 23 June 2023) <https://arstechnica.com/tech-policy/2023/06/lawyers-have-real-bad-day-in-court-after-citing-fake-cases-made-up-by-chatgpt/>accessed 13 March 2024. [1]Michalis Mountantonakis and Yannis Tzitzikas, ‘Validating ChatGPT through RDF Knowledge Graphs and Sentence Similarity’ (ACM Conference, OpenProceedings.org, July 2017)<https://arxiv.org/pdf/2311.04524.pdf> accessed 13 March 2024.
Michalis Mountantonakis and Yannis Tzitzikas, ‘Validating ChatGPT through RDF Knowledge Graphs and Sentence Similarity’ (ACM Conference, OpenProceedings.org, July 2017)<https://arxiv.org/pdf/2311.04524.pdf> accessed 13 March 2024.

  • Multi-layered legal system: Central legislation coexists alongside state-specific laws and diverse customary practices, creating a complex tapestry of legal sources.
  • Sheer volume of information: The sheer volume of legal information, encompassing vast legislative enactments, judicial pronouncements, and commentaries, makes it overwhelming to search and analyze manually.
  • Inconsistency in terminology: Legal language can vary across different jurisdictions, and courts leading to inconsistencies in the way legal concepts are expressed.


Apart from these, there are also variations in terms of practice, particularly noticeable in lower courts even within the same state. For instance, from practical experience, we have observed discrepancies in the Vakkalath template between different courts, such as those in Ernakulam and Chengannur in Kerala. This absence of standardization not only affects the efficiency of the system but also poses challenges to the digitisation process. The need for standardization becomes increasingly apparent, as it not only streamlines operations but also aids in the seamless integration of technology within the legal framework.

The adoption of Legal Ontologies (LOs) and Knowledge Graphs (KGs) offers a promising solution to these challenges. By capturing the precise meaning and relationships between legal concepts, LO/KGs can provide a structured and standardised foundation for legal information.

Fig 1

Challenges pertaining to interlinking of legal documents

Legal information systems facilitates the search for the required information for solving a legal problem. In India with the digitization of courts, courts are building their digital databases. The Supreme Court of India with the most sophisticated legal information system e-SCR compared to lower courts, provides a keyword-based search interface. Additional filters can be used to restrict the search, for example, a particular publication date of the judgment. The search results are then presented as a long result requiring users to go through all individual judgments and check whether they contain the required information. Furthermore, the documents are only partly interlinked, for instance, precedents referred to in the decision are not linked to the actual judgment. This requires users to start an additional search for each precedent referred. As a result, the missing links in the documents make the search more difficult to navigate and needlessly time-consuming. When judgments of lower courts/foreign courts are referred, the situation gets worse. The lower courts/foreign court’s legal information systems need to be consulted in such situations, which may be set up quite differently. The issue can be resolved by adding the missing links between the documents. Using a machine-readable data format like Resource Description Framework, it is possible to link data to enable structured queries and easier navigation through interlinked legal documents.

In 2011, the EU proposed standards for easing the interlinking of legal information across the EU member states based on RDF. The European Law Identifier (ELI) for legislative documents and European Case Law Identifier (ECLI) for judiciary documents assigns unique identifiers to describe a minimum set of metadata for legal documents, a starting point to overcome the disadvantages of the search process outlined above in the EU Member States.


Erwin Filtz, ‘Building and Processing a Knowledge-Graph for Legal Data’ in Evaa Blomqvist and others (eds), The Semantic Web, vol 10250 (Lecture Notes in Computer Science, Springer, 2017)  <https://link.springer.com/chapter/10.1007/978-3-319-58451-5_13#Sec3> accessed 13 March 2024.
Padmakshi Sharma, ‘CJI DY Chandrachud Launches E-SCR Project For Digital Version Of Judgements Reported In Supreme Court Reports’ ( LiveLaw, 2 January 2023)<https://www.livelaw.in/top-stories/cji-dy-chandrachud-launches-the-e-scr-project-for-digital-version-of-judgments-reported-in-supreme-court-reports-217859?code=6EpasU7aqWBB6QwnyRXuCW4x9U3j1v#> accessed 13 March 2024.

Erwin Filtz, ‘Knowledge Graphs for Analyzing and Searching Data’ (PhD Dissertation, Vienna University of Economics and Business 2021)<https://penni.wu.ac.at/supervision/Erwin%20Filtz%20Thesis%202021.pdf> accessed 13 March 2024.

Another big concern when it comes to legal research in India is the inconsistencies with the usage of legal terminologies – especially with the lower judiciary. Different formats are used in different courts for filing the same lawsuit, different languages are used in different courts . If someone needs to see the case status or a judgment in a specific case, there is a lack of consistency in terminologies for the same type of litigation. The difficulty increases when it comes to finding the judgements for the same type of litigation in the jurisdiction of other states. The sheer volume of legal information, coupled with inconsistencies and ambiguities across jurisdictions, creates significant barriers for both legal professionals and the public at large. Consequently, legal professionals and the public face difficulties in identifying relevant legal materials, understanding their context, and drawing accurate legal conclusions.

Part II

  

LO/KGs offer a compelling solution to the challenges plaguing India’s legal landscape. An ontology is a formal representation of a domain, capturing the key concepts, their attributes, and the relationships between them. A knowledge graph builds upon this foundation, creating a network of interconnected entities and their relationships, enabling efficient information retrieval and analysis.

The concept of ‘Knowledge Graph’ rose to prominence in 2012 when Google inculcated Knowledge Graph in its search engine which enhances search results by providing contextual information. It’s not just Google, major companies like eBay, Bloomberg, and Facebook all have adopted Knowledge Graphs to capture and analyze the semantic relationships between concepts, which benefits applications like personalised systems and knowledge bases.

Ajay Suman, ‘Lack of procedural uniformity in the Indian justice delivery system’ (Times of India, 5 June 2022)<https://timesofindia.indiatimes.com/readersblog/ajayamitabhsumanspeaks/lack-of-procedural-uniformity-in-the-indian-justice-delivery-system-43230/> accessed 13 March 2024.
Amit Singhal, ‘Introducing the Knowledge Graph: things, not strings’(Google, 16 May 2012) <https://blog.google/products/search/introducing-knowledge-graph-things-not/> accessed 13 March 2024.

Knowledge Graphs are not alien to the legal domain, initiatives like EUR-Lex, aiming to standardize legal documents in Europe; and project Lynx, which focuses on addressing challenges in legally structured data, have ingrained knowledge graphs into their operation.
In the context of the Indian legal system, LO/KGs can be designed to:

  • Standardise legal terminology: By capturing the precise meaning and relationships between legal concepts , LO/KGs can facilitate consistent interpretation and application of the law across different jurisdictions.
  • Organize legal information: By structuring legal data in a machine-readable format, LO/KGs enable efficient search and retrieval of relevant legal materials based on specific queries and contextual factors.
  • Reasoning ability to analyze legal information: LO/KGs can be equipped with reasoning capabilities, allowing them to identify potential conflicts, inconsistencies, and logical implications within the legal framework.

These capabilities of LO/KGs hold immense potential to bridge the divides within India’s legal system by:

  • Enhancing legal research: LO/KGs can empower legal professionals to conduct comprehensive and efficient research, saving valuable time and resources. They can provide context-sensitive information retrieval, identify relevant precedents, and highlight potential legal arguments.
  • Promoting legal awareness: By making legal information more accessible and understandable, LO/KGs can empower the public to better understand their rights and obligations under the law. This can contribute to a more informed and engaged citizenry.
  • Supporting legal innovation: LO/KGs can serve as the foundation for developing innovative legal applications, such as chatbots and decision-support systems, that can provide personalized legal guidance and assistance to individuals and businesses.

While the potential of LO/KGs is undeniable, it is crucial to recognize the unique characteristics of India’s legal system when developing these tools. This necessitates a collaborative effort involving legal experts, knowledge engineers, and technologists to:

  • Capture the nuances of Indian legal terminology: LO/KGs must be built upon a comprehensive understanding of legal concepts and their interpretation within the Indian context, incorporating regional variations and customary practices.
  • Integrate diverse legal sources: The knowledge graphs should encompass not only central legislation and judicial precedents but also state-specific laws, customary practices, and relevant commentaries.
  • Ensure accessibility and inclusivity: LO/KGs should be developed with accessibility in mind, utilizing vernacular languages and incorporating features that cater to individuals with disabilities.

Developing and deploying domain-specific LO/KGs tailored to India’s legal system holds immense potential to bridge the divides, enhance legal research, and contribute to a more robust and inclusive legal infrastructure. By fostering collaboration between legal professionals, knowledge engineers, and technologists, India can harness the power of LO/KGs to build a more robust and democratic legal infrastructure.