India’s Supreme Court Sets AI Rules for Courts, Invites Public Comments

India's Supreme Court released draft regulations on June 3, 2026, governing how AI tools can be used across courts, with a public comment deadline of June 20.

The regulations, prepared by the Supreme Court's AI Committee, apply to every court in India, from the Supreme Court down to subordinate courts, tribunals, and statutory commissions. The framework permits AI for legal research, drafting, translation, transcription, and case management. It bars AI from deciding cases, passing sentences, or assessing bail eligibility.

This matters because India has one of the world's largest judicial backlogs, and these regulations set the first formal national framework defining what AI can and cannot do inside courtrooms.

What AI Can Be Used For

The draft lists over a dozen approved uses for AI in court processes:

• Legal research and precedent retrieval

• Citation verification

• Summarising pleadings and judgments

• Translating legal documents, orders, and pleadings

• Automated transcription of court proceedings, subject to mandatory review by a designated officer

• Drafting assistance and generation of prescribed formats

• Preparation of notices, summons, and administrative documents

• Cause list preparation and hearing scheduling

• Case filing assistance and defect scrutiny

• Record management and judicial resource allocation

• AI-powered chatbots to help litigants access court services

• Accessibility tools such as text-to-speech, speech-to-text, and Braille translation for persons with disabilities

• Document authenticity verification and fraud detection in administrative processes, subject to mandatory human review

What AI Cannot Be Used For

The regulations list nine absolute prohibitions. No authority under the regulations, including the Apex Body, can relax these:

• Deciding cases or reaching judicial outcomes

• Passing sentences

• Determining bail eligibility

• Assessing flight risk or the likelihood of reoffending

• Evaluating witness credibility

• Predicting or profiling the future conduct of parties, accused persons, witnesses, or lawyers

• Conducting surveillance of judges, advocates, or litigants within or in connection with court premises

• Submitting AI-generated output as an independent source of evidence without full disclosure

• Using opaque black-box AI systems in matters affecting personal liberty

The draft specifically bans risk-scoring systems, defined as AI tools that assign a numerical or categorical score to estimate the probability of a person committing an offence, reoffending, or failing to appear before a court.

Disclosure and Accountability

Lawyers and litigants who use AI to prepare pleadings, documents, or evidence must declare this to the court at the time of submission. Regulation 43(3) requires the declaration to be made through a duly executed certificate in a prescribed format.

Courts can require additional information, including the identity of the AI system used, the extent of AI assistance provided, and the steps taken to verify the accuracy of AI-generated content.

Accountability rests entirely with the person submitting the material. Regulation 43(6) states that if any document submitted to a court is found to be fabricated, false, or misleading because of its AI-generated character, the person submitting it bears full responsibility. That person cannot use the AI-generated nature of the output as a defence.

Regulation 8 reinforces this: "It shall not be permissible to invoke the outputs of an AI System, the opaqueness of a Black Box system, or the occurrence of hallucination, as a ground for avoiding accountability for a palpably incorrect, illegal, or harmful decision."

The regulations also acknowledge the problem of AI hallucinations directly, defining them as instances where AI systems generate outputs that appear plausible but are factually incorrect, fabricated, or unsupported by verifiable data. All AI-generated outputs are treated as advisory in nature and must be verified before use.

What It Means for Legal Tech Vendors

The draft places significant compliance obligations on private technology companies that supply AI systems to courts.

No private entity can participate in any AI system deployed in court processes without prior written approval from the competent authority. All vendor contracts must include provisions on court data ownership, restrictions on using judicial data beyond the scope of the engagement, audit rights for the AI Secretariat, cybersecurity safeguards, and liability for AI-related harm.

Vendors must provide complete technical documentation of the architecture and training data of their AI systems. For high-risk AI tools, explainability documentation is mandatory. Contracts must also prohibit the retraining, fine-tuning, or modification of AI models using court data without express written approval from the AI Committee.

Where AI tools are developed using court data or court resources, the court retains ownership of, or a perpetual royalty-free licence to, the resulting tool and its outputs. No private entity can claim exclusive intellectual property rights over tools built primarily using judicial data or public resources.

Sensitive judicial data cannot be processed on external servers without express authorisation. Regulation 46 requires on-premise or sovereign cloud deployment for AI systems that process sensitive judicial data, which narrows the options for vendors whose products rely on third-party cloud infrastructure.

Any data breach or AI incident involving a vendor-supplied system must be reported to the court authority without delay. Non-compliance can result in suspension or termination of the engagement.

New Oversight Bodies

The draft proposes a permanent Apex Body at the Supreme Court level to govern AI use across the judiciary. Its members will include two Supreme Court judges, two Chief Justices of High Courts, two High Court judges, a technology expert, a cybersecurity specialist, a finance expert, a MeitY government officer, advocates with expertise in technology law, and the AI professor from the National Judicial Academy in Bhopal.

Every High Court will also establish an AI Committee and a dedicated AI Secretariat headed by an officer of District Judge rank.

A Centre of Research and Excellence on Artificial Intelligence, referred to as CoRE-AI, will conduct research, evaluate AI tools, and provide technical support to the judiciary.

All AI systems deployed in courts must undergo technical, legal, and ethical audits at intervals not exceeding one year. Courts must maintain an AI Register documenting every approved AI tool, its purpose, audit findings, and any incidents. An AI Incident Database will record malfunctions, errors, bias, and security breaches.

Every High Court, tribunal, and commission must publish an annual transparency report on AI adoption, listing systems in use, audit outcomes, and compliance measures.

The regulations also require that where AI tools are developed using court data or court resources, the court retains ownership of, or a perpetual royalty-free licence to, the resulting tool. No private entity can claim exclusive intellectual property rights over tools built primarily using judicial data.

Stakeholders and members of the public can submit comments and suggestions to the Member Secretary, AI Committee, Supreme Court of India, at office.regcc@sci.nic.in by June 20, 2026.