Author: Lasha, a student, Janhit college of law
To the point
Artificial intelligence in judicial context refers to computational system capable of document summarise, legal analzise, translation, transcription ,legal research and drafting assistance. AI is directly touched and spread in or around courts life, liberty, not as ai used in commerse and entertainment. Here question arise that whether the ai should be kept out the courtrooms or where the line between assistance and adjudication must be drawn.
Globally,regulators have cover this thing with single guiding idea: AI may inform a judicial decision but it must never becomes decision maker. The European union’s ai act explicitly classifies systems used by judicial authorities to research facts, interpret law. In united states the courts have grapple with risk assessment tools used in sentencing bail and even judges citing fabricated case law generated by large languages.
India have taken assistive approach: the supreme court is owe ai tools are designed to support research and translation ,but never to decide the cases.
The regulation of artificial intelligence in courts focuses heavily on the rule of “Human primacy,’’ ensuring technology serves only as assistive tool while leaving final legal and judicial authority entirely to humans.
In India, the framework is taking definitive, formal shape via the newly proposed REGULATION FOR USE OF ARTIFICIAL INTELLIGENCE IN COURTS,2026.
Abstract
Artificial intelligence has moved from the margins of legal technology into the heart of judicial administration. Courts across the world, including the supreme court of India, now used AI tools for legal research, translation, and transcription ,while litigants and even some judicial officers use generative AI for drafting and research.
This article examines the constitutional, statutory, and institutional framework that governs the use of AI in courts, surveys landmark judicial responses from india. It concludes that AI in courts must remain strictly assistive ,with human judges retaining final decision-making authority, supported by clear regulatory guardrails, disclosure norms and evidentiary safeguards.
To Proof
The clearest evidence of why AI in courts needs careful regulation comes from how courts themselves have already used and misused ,these tools.
The supreme court of India launched the supreme court portal for assistance of artificial intelligence in courts efficiency (SUPACE) in April 2021, it is designed to help judges and research staff sift through voluminous case records, identify relevant facts and locate precedents, without itself rendering any decision.
The vidhik anuvaad software( SUVAS) has translated tens and thousands of Supreme court judgements from English into Hindi and other regional languages.
The supreme court’s AI committee, originally constituted in 2019 and reconstituted in December 2025 under a sitting judge, has consistently drawn a firm line: such tools may assist with research, translation, and help to interpreted the law but AI cannot draft the judgement or predict outcomes.
Kerala’s judicial AI policy goes further, expressly prohibiting generative AI from any role in judicial decision making.
The Manipur high court was reported to have used chatGpt to work on bail jurisprudence.
Karnataka trial court ended up citing a precedent that, on closer scrutiny, had not actually been argued by either party,
An unmistakable AI hallucination finding its way into judicial reasoning.
These episodes are not isolated to india. Since 2023, courts across the united states have sanctioned lawyers, and in some cases entire law firms, for submitting briefs containing AI generated citiations to cases that simply do not exist. Monetary fines, referrals to bar disciplinary bodies, and mandatory continuing-legal-education orders have become standard judicial responses. The proof, in short, lies not in abstract risk forecasting but in a growing docket of real orders, real sanctions, and real reputational damage to the legal profession.
Case laws
1. Mata v. Aviance, Inc.(S.D.N.Y.,2023)
In this American case, attorneys submitted a legal brief containing several judicial opinions and citiations generated by chatGPT. When the fabrications were discovered, the court imposed monetary sanctions and required the lawyers to notify the judges whose names had been falsely attached to the invented decisions.
This is case is now treated at international court as foundational illustration of unverified generative AI in litigation.
2.ANI Media v. openAI Inc. (Delhi court):
A landmark Indian case where ANI Media contested that OpenAI used its copyrighted news content to train ChatGPT without authorization , raising foundational questions about copyright infringement, fair use, and jurisdictional applicability.
3.Thaler v. U.S. copyright office (USA):
US Supreme court denied certiorari, effectively cementing the rule that an AI cannot be named as an inventor or author on copyright and patent registrations. Similar cases, like the rejection of the “suryast” (RAGHAV AI) copyright registration in the US and the DABUS patent cases, have set global precedents that purely AI-generative outputs bear no copyright protection.
4.Anil kapoor v. Dilip khabani (delhi high court,2023):
Set foundation for the 2025-2026 cases. The court barred the digital morphing of the actors’s face and iconic catchphrases, establishing that “free speech” cannot shield commercial AI exploitation.
5.UMG Recordings, Inc. V. Suno, Inc. And UMG v. Udio (USA):
Major music groups (universal, sony) filed landmark lawsuits alleging massive copyright infringement by Suno and Udio.Record labels presented evidence that users could prompt the platforms to generated audio tracks that closely mirror protected, iconic sound recordings.The AI companies claim their falls under “fair use” as it represents transformative learning.
Use of legal jargon
The regalution of AI in courts sits at the intersection of constitutional law, evidence law administrative oversight. Article 21 of the constitution guarantees te right to life and personal liberty, which can be curtailed only through a procedure that is just, fair and reasonable. AI tool’s influence the decision of bail , sentencing or drafting , the submission is shape the courts reasoning , the demand for explainability it becomes constitutional not requirement.
Under the bharatiya sakshya adhiniyam (BSA), 2023 the defination of “document” in section 2(1)(d) expressly extends to electronic and digital records, meaning AI generated translation and transcription and even AI assisted drafts can themselves become digital document, can be challenged in the court proceedings.The parties must retain the right to test the reliability as the would test any other document evidence because the AI generative records or documents might be un-accurate that reflect the proceeding of the court.
Internationally through instruments like EU AI Act, which classifies judicial research AI as “high risk” and insist of that the final decision making must be remain in the authority of human judge and justice.
In India, the AI governance guidelines released by ministry of electronic and information technology in November 2025, it adopts “techno legal” which is principal based model rather than a rigid statute, relying on the existing laws such as the information technology act,2000 and digital personal data protection act,2023.
Doctrine of natural justice provide right to fair hearing, unbiased justice. A litigant has the right to know on which basis the decision is affected and to challenge the basis.
Conclusion
Algorithms are explicitly barred from adjudicating cases, evaluating the witness credibility, or determining bail, sentencing . The “human in the loop” ia non negotiable.
Because of high risk decisions remain human only, AI is highly valued for docket prioritization, automated transcriptions, legal research, and backlogs.
The future of the legal system will be determined not merely by how advanced these tools become but how the courts strictly enforce the humansight.
The integration of AI in the judicial system is no longer futuristic theory but an active reality. Court globally – including supreme court of India with its proposed drafting regulations –are actively defining the boundaries of legal technology.
FAQs
Q.1.Does any Indian law currently regulate the use of AI in courts?
No dedicated statue exist yet. Use is governed by article 21,the IT ACT,2000 and DPDP,2023,the BSA’s electronic records provisions and AI governance guidelines.
Q.2.Can a lawyer be penalized for al generated fake citations?
Yes, Indian and foreign courts have dismissed pleadings and imposed sanctions, contempt notices, or bar referrals for unverified AI generated citations
