 (4).png)
1 week ago
4
ARTICLE AD BOX
Summary
As fake citations and AI hallucinations creep into Indian courtrooms, the apex court is drawing a strict line between human judgement and machine assistance.
On 3 June the Supreme Court released draft regulations proposing a comprehensive framework for the adoption of artificial intelligence (AI) in India’s judicial system, marking its first-ever attempt to formally regulate AI’s use in India's judiciary. It has invited public comments until 20 June.
The proposal rests on a simple core principle: AI may assist judicial work, but decision-making must remain exclusively with humans.
Mint takes a close look at the proposed guidelines and what they could mean for the future Indian courts.
Why did the court issue these draft rules?
The Supreme Court introduced the draft regulations to address the rapid integration of AI tools within the legal profession. The guidelines aim to establish clear boundaries as concerns mount over the potential misuse of these technologies.
The move follows several instances of AI hallucinations—incorrect or fabricated AI-generated content—being mistakenly included in court proceedings.
- In February 2025, an order of the Income Tax Appellate Tribunal (ITAT) in Bengaluru was withdrawn after it cited non-existent judicial precedents.
- A month later, the Karnataka High Court called for action against a civil court judge after an order referred to Supreme Court and Delhi High Court judgments that did not exist.
- In March 2026, the Supreme Court took note of allegations that the NCLT and NCLAT had relied on allegedly fake or non-existent judicial citations in insolvency proceedings involving Essel Infraprojects Ltd.
The Supreme Court constituted its first AI committee in 2019 to explore how the technology could be integrated responsibly into the justice system. In December 2025, Chief Justice of India Surya Kant reconstituted the AI committee and tasked it with steering the adoption, development and deployment of AI tools across the higher judiciary and subordinate courts. The committee is chaired by Justice P.S. Narasimha, and includes several high court judges, technology experts and judicial administrators.
Where is AI allowed?
The committee's proposal strongly favours the responsible adoption of AI in courts. Under the draft framework, AI can be used for legal research, precedent retrieval, citation verification, translating judgments and pleadings, transcripting court proceedings, summarizing documents, managing cases, scheduling, court administration and backlog monitoring.
The regulations also permit the use of AI chatbots to assist litigants, accessibility tools for people with disabilities, anonymisation of court records, and verifying that documents are authentic. However, all AI-generated output will remain advisory and subject to human review.
Where is it not allowed?
The committee draws a line around judicial decision-making. The draft regulations prohibit AI from deciding cases, determining bail, imposing sentences, assessing the credibility of witnesses or litigants, predicting judicial outcomes, or assigning risk scores to accused persons. The framework also bars AI from predicting future behaviour, such as whether a person is likely to reoffend or abscond.
Also, opaque AI systems cannot be used in matters affecting legal rights or personal liberty. AI-based surveillance of judges, lawyers or litigants is also prohibited, as is the use of personal data for training AI systems without proper authorisation. Any AI-generated material submitted before a court must be disclosed as such and verified.
The draft guidelines also propose a grievance redressal mechanism, allowing litigants who suffer harm due to the improper or prohibited use of AI in court proceedings to approach the concerned court for relief.
How will AI be deployed?
The draft proposes a tightly regulated deployment process. Before an AI tool is adopted, it must undergo a technical and ethical impact assessment that examines its accuracy, bias, explainability, privacy protections and cybersecurity safeguards.
AI systems must also comply with the Digital Personal Data Protection Act, 2023, and other laws governing judicial and personal data. Courts may also test these tools in a controlled environment before they are deployed at scale. Once approved, AI tools will remain subject to human oversight, with judges and court officials retaining final authority over all decisions.
The framework also mandates annual technical, legal, ethical and cybersecurity audits, as well as incident reporting and an AI register to document approved systems. Private AI vendors can be engaged only with prior approval and must comply with strict requirements on data protection, transparency and accountability.
Every court will also be required to establish emergency and fall-back protocols to ensure operations continue even if an AI system fails or is suspended.
How could AI change judicial work?
If implemented effectively, AI could transform judicial administration by handling routine tasks such as legal research, document review, translation, transcription and case management, allowing judges to focus on adjudication and legal reasoning.
It could help reduce case backlogs, improve efficiency through smarter scheduling and automated workflows, and make court services more accessible through translation tools and digital assistants.
Who will oversee the use of AI in courts?
The committee has proposed a dedicated governance framework to oversee AI adoption. At the top will be a Supreme Court-led apex body, responsible for setting standards, approving AI systems, and coordinating policy across the country.
The proposed framework also proposes AI committees in the Supreme Court and every high court, supported by dedicated AI secretariats tasked with approvals, audits, incident reporting and compliance monitoring. It also envisages a Centre of Research and Excellence on Artificial Intelligence (CoRE-AI), which will conduct research, evaluate emerging technologies, and provide technical support to the judiciary. Together, these institutions will oversee audits, cybersecurity, grievance redressal and policy development.
What do experts say about the proposed rules?
Mihir Rale, partner and co-head, digital+, TMT at Cyril Amarchand Mangaldas, said the proposal’s significance is clear given India's massive legal backlog. With 55.8 million pending cases, a severe shortage of judges, and deep recruitment challenges, clearing the backlog under the current system could take over 300 years, he said. "The only real option is to enhance productivity, and the adoption of AI is certainly a game-changer in this context," he added.
Vaibhavi Sharma, associate partner at Prosoll Law, said, “Over the next 5 to 10 years, AI is likely to transform the judiciary through improved legal research, real-time transcription, multilingual translation, intelligent case management, backlog monitoring and enhanced access to justice.”
However, lawyers cautioned that implementation would be a key challenge. Ankit Sahni, partner at Ajay Sahni Associates, said courts will need institutional capacity to audit systems, manage vendors, and safeguard sensitive data. He also highlighted that consistency across jurisdictions could be difficult, as courts may adopt infrastructure at varying speeds. "Unless there are clear standards, sufficient funding and continuous capacity-building, compliance may be uneven," he added.
How have Indian courts system used AI so far?
Since the Supreme Court constituted its first AI committee in 2019, the judiciary has introduced several AI-driven initiatives. Among the most notable of these was the Supreme Court Portal for Assistance in Court Efficiency (SUPACE), an AI research tool launched in April 2021 and designed to help judges process case records and identify relevant precedents. In November 2019 the court launched the Supreme Court Vidhik Anuvaad Software (SUVAS), which uses AI to translate judgements into several Indian languages.
About the Author
Krishna Yadav
Krishna Yadav is a Senior Correspondent at Mint, based in New Delhi, and part of the corporate bureau. He joined the newsroom as a trainee in 2023 and quickly grew into his current role. He writes on legal and regulatory developments in corporate India, with a focus on insolvency, taxation, company law, and policy. His reporting includes tracking and breaking key legal stories from the Supreme Court, Delhi High Court, NCLT, and NCLAT.<br><br>With a background in law, Krishna is known for simplifying complex legal developments into clear, accessible stories for readers. His work focuses on trends in corporate law and policy that affect businesses. This ranges from explaining tax disputes—like whether coconut hair oil is edible—to writing on why celebrities are seeking personal rights protection. He closely tracks India’s insolvency system, covering issues such as creditor losses, gaps in the process, and challenges in how the framework works in practice.<br><br>Krishna also tracks developments within law firms—covering hiring trends, how firms help companies navigate global challenges, and how the legal industry is adapting to artificial intelligence. Beyond legal reporting, he has written long-form pieces, including on-ground coverage of the 2024 general elections, capturing the scale and logistics of polling across India.<br><br>Outside work, he enjoys travelling, exploring new places, and reading about geopolitics and history.
English (US) ·