INTRODUCTION
Academic discussions about ‘law and technology’ as a subject that cuts across each other have already existed for a long time. The rise of Artificial Intelligence (AI) nonetheless adds a new chapter to the story. Justice AI is about to revolutionise the way legal judgments are made. From legal research to judging cases and, increasingly, to forecasting verdicts, AI is rapidly becoming a means to expedite justice. But this shift also raises important ethical, responsible, transparent, and unbiased questions.”
In India, where delays and access to justice are significant concerns, the domain of AI for court functioning is both fascinating and complex. But AI is a computer system that can perform tasks that normally require human intelligence. In the legal context, AI includes natural language processing, machine learning, and data analytics tools used by lawyers and judges for decision-making. The impact of these weapons, especially in the precincts of the Indian judiciary, is ripe for scrutiny.
APPLICATIONS OF AI IN JUDICIAL DECISION-MAKING
Nations around the world, including the U.S. and China, have already incorporated artificial intelligence into their justice systems. In the U.S., the COMPAS tool helps judges by offering risk assessments for granting bail, but it has been accused of racial bias and opacity. [1] AI is used to predict sentencing and contradictions in testimony in China through the ‘206 System’, which is an AI-powered judicial tool used in China to assist in criminal trials. It analyses case files, checks evidence validity, and flags procedural errors before trial. The system also suggests sentencing based on past precedents and legal guidelines. Integrated with surveillance and police databases, it aims to standardise and streamline judicial decisions. [2]
India, too, has made important progress. As of 2021, the Supreme Court of India has launched SUPACE (Supreme Court Portal for Assistance in Court Efficiency), which uses AI to help judges with legal research, case summaries, and identifying case law.
Importantly, it was also explained that SUPACE would not be providing evidence so much as supporting judges to help them manage their caseloads. [3]
ADVANTAGES OF AI IN THE JUSTICE SYSTEM
- EFFICIENCY: When AI can handle routine legal work, like filing (sort of) and citation analysis, and templates for drafting, judges and lawyers can do the work of judging and lawyering.
- THE HASTENING OF PENDENCY: In addition to helping in scheduling, prioritising cases, and maintaining the cause-list, AI tools can be of great help in drastically reducing the pendency of 4.5 crore+ pending cases in India. [4]
- CONSISTENCY: AI can spot similar precedents and statutory interpretations in other jurisdictions, thereby also mitigating discrepancies in decisions on similar facts.
- ACCESS TO JUSTICE: AI-powered platforms can democratize access to law to the underprivileged and geographically isolated segments of our society, more so through vernacular translation. [5]
- POLICYMAKING: Judicial analytics can lead to law reform by uncovering the pattern of delays, differences in regional variances, or systemic inefficiencies.
CHALLENGES AND ETHICAL CONCERNS
Yet while AI in judicial decision-making has obvious advantages, it also holds huge downsides.
First, THE TRANSPARENCY AND ‘BLACK BOX’ ISSUE: Most AI algorithms, especially the ones that leverage deep learning, are not explainable. The logic behind such reasoning orders is particularly important in the judicial realm because it is the lifeblood of the process. If a tool is unable to explain why it made a recommendation, it raises due process concerns. [6]
Second, BIAS IN TRAINING DATA: –AI systems learn from historical data. If such data includes past discriminatory practices—say, harsher sentencing of Dalits or Muslims—the AI might replicate and amplify those biases. This has already been demonstrated in international contexts like COMPAS. [7]
Third ACCOUNTABILITY AND LIABILITY: – A central legal question is: Who is liable if an AI-influenced decision causes harm? Is it the judge, the developer, or the institution? Indian jurisprudence is yet to address this grey area, raising concerns about accountability. [8]
Fourth, DATA PRIVACY AND SECURITY: – AI tools require access to vast databases, court records, personal details, and sensitive information. While the Digital Personal Data Protection Act of 2023 lays down a framework for data governance, there are concerns about the security of judicial data.[9]
Fifth, The Infrastructure Gaps: – Many lower courts, especially in rural India, lack reliable internet connectivity or digitised records. AI implementation will widen the urban-rural divide unless these systemic gaps are addressed. [10]
JUDICIAL TRAINING AND AI LITERACY
To implement AI in courts, one must educate judicial officers about AI. The National Judicial Academy has started conducting workshops and training programs on the utilisation of AI tools. We are also working in partnership with institutes such as IIT Delhi & IIT Hyderabad to develop explainable AI models tailored to the Indian Legal domain. [11]
Without proper comprehension, over-dependence or misuse of AI tools is possible. Judges must therefore be the ultimate decision-makers, and AI is an assistive mechanism rather than a substitute.
TRADITIONAL ADJUDICATION VS. AI-ENABLED JUDGING
In the more traditional adjudication, the judge is not just a lawyer but a moralist with the capacity for empathy in weighing human situations. AI is instead data-driven; it is devoid of the human touch that you need in family law, juvenile justice, or restorative sentencing.
For example, in the case of domestic violence, AI might flag incidents based on keywords, but it cannot understand subtleties like fear, trauma, or manipulation. So, we should never be ruled by AI, only ruled by the AI. [12]
REGULATORY FRAMEWORK IN INDIA
There is no exclusive law governing AI in the adjudication process in India at the moment. The National Strategy for AI (2018) by NITI Aayog has a section on ethical AI, but it is not a binding guideline. [13]
Legal expertise indicates the importance of a multilayer system of governance, inter alia:
- Independent ethics committees to review AI tools that are to be used in courts.
- Compulsory audits on bias and accuracy.
- Open-source algorithms for increased transparency.
- Binding usage policies for data that are in line with privacy rights.
Additionally, the judiciary could adopt a Code of AI Ethics, much like the Bangalore Principles for judicial conduct, tailored to regulate AI usage.
COMPARATIVE GLOBAL PERSPECTIVES
- Estonia is working on an A. I “judge” for disputes below €7,000, which is efficient, but raises the potential problem of judicial impersonality. [14]
- China’s Smart Courts, which feature facial recognition, automated document verification, and virtual trials, are praised but also slammed as surveillance and dissent-free. [15]
- Brazil has created ‘Victor’—an AI assistant to help the country’s Supreme Federal Court filter cases for constitutional relevance. [16]
India has to calibrate the fine balance between technological advances and constitutional values, particularly Article 21 [17] (right to a fair trial) and Article 14 [18] (equality before the law).
CONCLUSION
The contribution of artificial intelligence to India’s judicial process is promising and perilous at the same time. Although it provides opportunities for efficiency, consistency, and evidence-based reform, it is also vulnerable to threats of obscurity, bias, and ethical misjudgment. The future of the Indian judiciary is in humanity and technology working in combination, not in defeating the former. But courts must make sure that technology serves justice, not trumps it. To that end, India needs a robust legal framework, judicial training modules, and participatory policymaking that includes legal experts, technologists, civil society, and judges. As we welcome the AI era, we must ensure it remains a servant of the law, not its master.
Author(s) Name: Anurag Kumar (Chanakya National Law University)
References:
[1] Julia Angwin et al, ‘Machine Bias’ ProPublica (23 May 2016) https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing accessed 8 June 2025.
[2] Xin He, ‘Smart Courts in China: Opportunities and Risks’ (2020) Asian Journal of Law and Society https://doi.org/10.1017/als.2020.6 accessed 8 June 2025.
[3] Press Trust of India, ‘SC Launches AI Tool SUPACE’ Hindustan Times (6 April 2021) https://www.hindustantimes.com accessed 9 June 2025.
[4] National Judicial Data Grid, ‘NJDG’ < https://njdg.ecourts.gov.in/njdg_v3/ > accessed 9 June 2025
[5] NITI Aayog, ‘Approach Document for India: Part 2 – Operationalizing Principles for Responsible AI’ (August 2021) < https://www.niti.gov.in/sites/default/files/2021-08/Part2-Responsible-AI-12082021.pdf > accessed 9 June 2025.
[6] Sandra Wachter et al, ‘Why a Right to Explanation of Automated Decision-Making Does Not Exist in the General Data Protection Regulation’ (2017) International Data Privacy Law 7(2).
[7] Vidushi Marda, ‘Artificial Intelligence Policy in India: A Framework for Ethical Governance’ (2018) Carnegie India https://carnegieindia.org accessed 10 June 2025.
[8] Ramesh Subramaniam, ‘Judicial Accountability in the Age of AI’ (2023) Indian Journal of Law & Tech 19(3).
[9] The Digital Personal Data Protection Act, 2023 (India).
[10] Ministry of Law and Justice, ‘e-Courts Phase III Vision Document’ (2022) https://ecommitteesci.gov.in accessed 10 June 2025.
[11] National Judicial Academy, ‘Training Calendar 2023–24’ https://nja.gov.in accessed 11 June 2025.
[12] Pratiksha Baxi, ‘The Human Element in Adjudication’ (2019) EPW Vol. 54, Issue 35.
[13] NITI Aayog, National Strategy for Artificial Intelligence (2018) https://niti.gov.in accessed 11 June 2025.
[14] BBC, ‘Estonia’s Robot Judge Will Help Settle Small Claims’ (25 Oct 2019) https://www.bbc.com/news accessed 11 June 2025.
[15] Xin He, ‘Smart Courts in China: Opportunities and Risks’ (2020) Asian Journal of Law and Society https://doi.org/10.1017/als.2020.6 accessed 11 June 2025.
[16] Supremo Tribunal Federal, Brazil, ‘Victor AI Project’ https://portal.stf.jus.br accessed 11 June 2025.
[17] The Constitution of India 1950, art 21
[18] The Constitution of India 1950, art 14
