IMPACT OF ARTIFICIAL INTELLIGENCE AND AUTOMATION ON LEGAL PRACTICE AND THE INDIAN JUDICIAL SYSTEM
Author Credit: Lara Borges & Zeba Darvesh - ANB Legal
Email: nazneen@anblegal.com
--------------
INTRODUCTION:
Artificial Intelligence (hereinafter referred to as ‘AI’) is increasingly being integrated into various sectors. With respect to its entry into the legal industry, it is enhancing the capabilities of legal professionals. AI tools, applications, and software improve efficiency and accuracy in document review, due diligence, contract management, and predictive analysis.
To discern the true meaning of AI, it is necessary to define its three constituent elements, which are- an algorithm, machine learning, and deep learning[1]. An ‘Algorithm’ is a set of automated instructions executed systematically at high speed to solve specific problems. By using algorithm-driven systems, it is possible to automate various decisions and tasks such as classifying, searching, scoring, ordering, ranking, selecting, and filtering data. Another key component of AI is ‘Machine Learning’, which involves analysing data to learn, predict, and make decisions based on various factors. Programs using it can learn to perform tasks correctly and improve over time with little or no explicit instructions on how to do those tasks. ‘Deep Learning’ is a machine learning technique that enables machines and autonomous systems to learn from examples. Instead of following pre-determined instructions, deep learning allows the machine to evaluate examples and identify patterns to solve future problems. By combining these components, AI can process instructions and deliver the desired outcomes.
REDEFINING LEGAL PROCEEDINGS AND BOUNDARIES:
Adopting AI in the legal industry involves ethical and regulatory challenges. AI algorithms can be prone to errors, and create ethical dilemmas, which can have serious consequences in legal matters. Issues like algorithmic bias, data privacy violations, and ethical misconduct highlight the need to address these concerns to ensure that AI is used responsibly in the legal field.
The integration of AI into the judicial system presents the ‘centaur’s dilemma[2],’ which addresses the degree of human control over AI. This concept highlights the predicament of AI’s influence. There is a quandary between reducing human control to let AI operate optimally without excessive interference or increasing oversight to ensure that outcomes are fair and reasonable.
Distinctively, the United States District Court for the Southern District of New York in Da SilvaMoore v. Publicis Groupe & MSL Groupe[3] held that computer-assisted review, or AI, could be “judicially approved for use” in appropriate cases.
Legal professionals often manage large volumes of documents during, inter alia, litigation, due diligence, and contract management. AI tools can efficiently be utilized to process vast amounts of such data, distinguishing pertinent information, thereby significantly reducing time and costs. They can also analyse legal precedents and provisions. Such tools are so advanced that they have the potential to provide historical data that helps legal professionals make informed decisions and develop litigation strategies, notably due to their machine learning technology.
Indubitably, no technology, regardless of its sophistication, can ever replace a human being. However, it can assist judges in their decision-making process, ensuring that handling a large volume of cases does not compromise justice. AI’s abilities in legal prediction, analysing past cases, and automating tasks like e-notices and e-summons help reduce the judicial system's workload. Additionally, AI has been crucial in facilitating virtual courts through e-courts and video conferencing, as demonstrated by the Supreme Court of India's Live Transcription Project.
INDIAN JUDICIARY’S COMMINGLING WITH AI:
The Chief Justice of India had expressed disquietude about the potential for social division and increased class discrimination regarding equal access to AI and its advancements[4]. Citing the renowned American freedom fighter Frederick Douglass, his Lordship had highlighted disquiet about AI disproportionately benefiting the wealthy at the expense of the poor, particularly in the pursuit of justice. There is a fear, as his Lordship remarked, that the adoption of AI might create a two-tiered system, where the poor could end up with substandard AI-driven assistance, while only the affluent or high-end law firms might fully leverage AI's transformative benefits. This scenario risks widening the justice gap and perpetuating existing inequalities in the legal system.
In March 2023, the Punjab & Haryana High Court utilized the AI tool ChatGPT in a matter concerning bail decision[5], thus steering the ship of AI into the legal arena. The bench, led by Justice Anoop Chitkara, was reviewing a bail petition for an individual arrested in June 2020 on various grave charges of rioting, criminal conspiracy, criminal intimidation, and murder. The court sought ChatGPT's opinion on global legal jurisprudence concerning bail in cases involving cruelty. After considering the input, the bench denied the bail application. This marked the first instance of ChatGPT being used to decide on a bail application in India. However, the Court had cautiously remarked that any mention of ChatGPT and the observations made are not to be taken as an opinion on the merits of the case, nor should the trial court consider these comments. Further, the High Court opined that such reference is solely meant to provide a broader perspective on bail jurisprudence when cruelty is involved. Indeed, it cannot be denied that the decision-making process of AI often seems opaque, posing significant concerns for the rule of law and justice. When AI operates as a “black box”, the true mechanisms behind its decisions are concealed from the user. This lack of transparency regarding the ratio decidendi of verdicts, particularly if AI is ostensibly involved in judicial decision-making, raises apprehension not only for judges but also for all the concerned parties.
The use of AI has been integrated into the Indian judiciary nascently. IIT-Kharagpur has developed an AI system designed to read legal orders[6]. Since 2021, the Supreme Court of India has employed AI-powered tools to manage information and provide it to judges for their decision-making. Although these tools do not make decisions themselves, they are instrumental in managing the large volumes of data involved in legal cases. One such instrument is the Supreme Court Vidhik Anuvaad Software[7] (“SUVAS”), which translates legal documents between English and vernacular languages, thereby enhancing the accessibility and efficiency of the legal process.
The National Judicial Data Grid (“NJDG”) is a key example of how technology, including AI, is being used to improve the legal system in India[8]. The Supreme Court aims to utilize machine learning through the Supreme Court Portal for Assistance in Court's Efficiency (“SUPACE”) portal to manage the extensive data received for various cases[9]. This hybrid system effectively combines human intelligence with machine learning, leveraging the strengths of both. The AI tool is designed solely to process information and provide it to judges for decision-making, without directly participating in the decision-making process. Judges handling criminal cases will use it on an experimental basis.
This support can help reduce the average trial duration, leading to greater efficiency in resolving cases and addressing the backlog in the legal system. Achieving this would fulfil the commendable goal of providing the public with efficient and enduring justice.
A COMPARATIVE ANALYSIS WITH USA:
The USA has adopted AI-powered tools like the Correctional Offender Management Profiling for Alternative Solutions[10] (“COMPAS”) to help judges with risk assessments by evaluating factors such as criminal history, socio-economic background, and mental health to predict the probability of reoffending. Additionally, the US Sentencing Commission employs AI to develop and implement sentencing guidelines, ensuring fair and equitable punishment. Moreover, the US court system uses chatbots to respond to frequently asked questions about court procedures, schedules, and other relevant topics, reducing the workload of court staff and making information more accessible to the public. A team of American scholars has also created a machine learning application that purportedly predicts Supreme Court of the United States case outcomes with 70.2% accuracy and forecasts individual justices' voting behaviour with 71.9% accuracy[11].
However, here arises the trouble, studies have found that the widely used COMPAS is biased against racial groups which implies that it has an algorithm that predicts a higher risk of reoffending for certain groups of individuals[12].
CONCLUSION:
As we incorporate AI into the Indian judicial landscape, it is crucial to actively tackle systemic hurdles and ensure that AI tools contribute to, rather than detract from, the quest for justice for all. Through embracing collaboration and encouraging global cooperation, a groundwork can be established for responsible and ethical promulgation of AI worldwide. This sets the stage for a future where technology empowers and benefits every individual, promoting inclusivity, innovation, and advancement.
[1] Xu, Y., Liu, X., Cao, X., Huang, C., Liu, E., Qian, S., Liu, X., Wu, Y., Dong, F., Qiu, C. W., Qiu, J., Hua, K., Su, W. Wu, J., Xu, H., Han, Y., Fu, C., Yin, Z., Liu, M., Zhang, J. (2021). Artificial intelligence: A powerful paradigm for scientific research. The Innovation, 2(4), 100179.
[2] Baker, J. E. (2021). The Centaur’s Dilemma: National Security Law for the Coming AI Revolution. Brookings Institution Press.
[3] 287 F.R.D. 182 (S.D.N.Y. 2012)
[4] Live: Conference on Technology and Dialogue between the Supreme Court of India and Singapore | Streamed Live on April 13, 2024 |India | ANI News
[5] Jaswinder Singh @ Jassi vs. State of Punjab and Another, 2024 PHHC 014669
[6] The KGP Chronicle: Official News Portal of the IIT KGP, January 16, 2020, Chirosree Basu
[7] Supreme Court of India, Press Release (2019, November 25).
[8] Policy, V. C. F. L. (2021, November 11). AI Innovation in Indian Judiciary a Distant Dream Without an Open Data Policy. Vidhi Centre for Legal Policy. https://vidhilegalpolicy.in/blog/ai-innovation-in-indian-judiciary-a-distant-dream-without-an-open-data-policy/
[9] Artificial Intelligence and Judicial Bias - Centre for Law & Policy Research. (2021, August 28). Centre for Law & Policy Research. https://clpr.org.in/blog/artificial-intelligence-and-the-courts/
[10] DOC COMPAS. (n.d.). https://doc.wi.gov/Pages/AboutDOC/COMPAS.aspx
[11] Katz, Daniel & II, Michael & Blackman, Josh. (2014). Predicting the Behaviour of the Supreme Court of the United States: A General Approach. SSRN Electronic Journal.
[12] Sam Corbett-Davies, Pierson E., Feller A. and Goel,S. (2016). "A computer program used for bail and sentencing decisions was labelled biased against blacks. It's actually not that clear". The Washington Post. Retrieved January 1, 2018.
Leave a comment