Supreme Court has released a draft set of regulations governing the use of Artificial Intelligence (AI) across the judiciary. This is a huge sign that AI will become part of the country’s justice delivery system. The proposed Regulations for Use of Artificial Intelligence in Courts, 2026 allow AI tools to assist judges, lawyers and court staff in a wide range of functions, from legal research and drafting to translation and case management. However, the draft draws a firm line that AI can assist, but it cannot decide. The regulations have been prepared by the Supreme Court’s Artificial Intelligence Committee and are open for public comments until June 20.
Human judges remain the final authority
The Supreme Court has made it clear that AI systems can only function in an assistive capacity and cannot replace judicial decision-making. The regulations state that AI must not supplant or compromise the independent exercise of judicial authority. The power to determine questions of law, facts and justice will remain exclusively with judges. Even if AI tools are used in the judicial process, accountability for decisions will continue to rest with judicial officers and court officials. Importantly, judges and court staff will not be allowed to defend incorrect or harmful decisions by blaming AI systems, algorithmic errors or AI-generated hallucinations.
Lawyers can use AI, but must disclose it
One of the most significant features of the draft regulations is the requirement that lawyers and litigants disclose when AI has been used in preparing court filings. If an AI tool is used to draft pleadings, prepare documents, generate evidence summaries or assist in submissions before a court, the user must inform the court at the time of filing. This disclosure will have to be made through a prescribed declaration or certificate. Courts may also ask parties to reveal which AI system was used, how extensively it was used and what steps were taken to verify the accuracy of the AI-generated content.
Users remain responsible for AI-generated mistakes
The draft regulations leave no room for shifting blame to technology. If a document, pleading or evidence submitted before a court is found to be fabricated, false, misleading or inaccurate because of AI-generated content, the person who submitted it will bear full responsibility. The regulations specifically state that a litigant or lawyer cannot use the AI-generated nature of the material as a defence. Courts will also have the power to pass appropriate orders against those responsible for such filings.
AI-generated evidence cannot be introduced secretly
The regulations require transparency whenever AI-generated material is used in legal proceedings. Any AI-generated information, synthetic content or AI-assisted evidence must be disclosed to the court. Such material cannot be presented as independent evidence without revealing its AI-generated character. This provision is aimed at preventing the misuse of increasingly sophisticated AI tools capable of generating realistic text, images, audio and video.
Courts must disclose significant AI use
The transparency requirements are not limited to lawyers and litigants. If courts use AI tools in ways that significantly influence case management, document analysis or judicial administration and such use could affect proceedings, the parties involved must be informed in a timely and accessible manner. The objective is to ensure that litigants are aware whenever AI plays a meaningful role in processes affecting their cases.
What AI will be allowed to do
The draft regulations permit AI to be used for a wide range of administrative and support functions within the judicial system. AI may assist in legal research, retrieval of precedents, citation verification, summarisation of pleadings and judgments, translation of legal documents and automated transcription of court proceedings.
It can also be used for drafting assistance, generating standard formats, preparing notices and summons, scheduling hearings, preparing cause lists, managing case records and conducting defect scrutiny. The regulations additionally encourage the use of AI-powered accessibility tools for persons with disabilities and AI chatbots that help litigants navigate court services.
What AI will never be allowed to do
AI systems will not be allowed to decide cases, determine judicial outcomes or pass sentences. They cannot decide bail applications, evaluate witness credibility or assess whether an accused person is likely to commit future offences. The regulations also prohibit AI systems from predicting the future behaviour of litigants, lawyers, witnesses or accused persons.
Courts will be barred from using AI-based risk-scoring systems that estimate the likelihood of reoffending, fleeing from justice or qualifying for bail. In addition, AI cannot interfere with judicial deliberations or rely on opaque “black-box” systems in matters affecting legal rights or personal liberty.
Surveillance through AI is prohibited
The draft regulations place clear limits on monitoring and surveillance. AI systems cannot be used for surveillance or continuous monitoring of judges, advocates, litigants or any other individuals within court premises. The Supreme Court has explicitly acknowledged one of the biggest challenges posed by generative AI: hallucinations. The regulations describe hallucinations as situations where AI systems generate information that appears plausible but is actually inaccurate, fabricated or misleading. This can include non-existent judicial precedents, incorrect legal provisions and false factual claims. To address this risk, AI-generated outputs must generally be verified before being relied upon, and such outputs are to be treated as advisory rather than authoritative.
A new governance structure for judicial AI
The draft regulations propose the creation of a permanent institutional framework to oversee AI deployment in the judiciary. At the national level, an apex body would be established under the Supreme Court to regulate and supervise AI adoption across courts. The body would include Supreme Court judges, High Court judges, technology experts, cybersecurity specialists, financial experts and advocates with expertise in technology law. Its responsibilities would include setting standards, approving AI systems, coordinating with High Courts and publishing annual governance reports. Separate AI Committees and AI Secretariats would also be established in the Supreme Court and every High Court to supervise implementation, investigate incidents and ensure compliance.
A dedicated AI research centre
The framework also proposes the creation of a Centre of Research and Excellence on Artificial Intelligence (CoRE-AI). The centre would conduct research, evaluate emerging AI technologies and provide technical support to the judiciary. It is expected to serve as a knowledge hub for courts as AI tools continue to evolve.
To ensure accountability, every AI system used within courts would be subject to technical, legal and ethical audits at least once every year. Courts would also be required to maintain an AI Register documenting approved AI systems, their intended uses, audit results, conditions attached to their deployment and any incidents associated with them. Alongside this, an AI Incident Database would be maintained to record malfunctions, bias-related issues, security breaches and other AI-related problems. Any malfunction or error that could have legal consequences would need to be reported immediately to the relevant AI Secretariat.
AI content verification authority proposed
The draft regulations propose the establishment of an AI Content Verification Authority. The authority would be responsible for creating standards, verification protocols and tools for examining AI-generated content used in court processes. This is intended to improve confidence in AI-assisted material appearing before courts.
Restrictions on private technology companies
The Supreme Court has also proposed safeguards around the involvement of private AI vendors. No private company will be allowed to participate in court AI systems without prior approval from the competent authority. Contracts with vendors would have to include provisions on data ownership, cybersecurity, audit rights, disclosure obligations and liability for AI-related harm.
Strong focus on data protection
The draft regulations require compliance with India’s data protection framework. Personal data cannot be used to train or improve AI systems without prior approval. Sensitive judicial information will receive additional protection and cannot be transferred to external systems without authorisation. The framework also mandates anonymisation wherever possible, adherence to data minimisation principles and regular cybersecurity audits.
Training judges and lawyers
Recognising that AI adoption requires institutional readiness, the regulations call for regular training programmes for judges, lawyers and court staff. These programmes would cover technical aspects of AI as well as issues such as algorithmic bias, hallucinations, data protection, cybersecurity risks and incident reporting mechanisms. Courts would also be required to establish emergency protocols and fallback systems to ensure that judicial operations can continue if AI tools fail or become unavailable.