Artificial intelligence is consistently becoming part of legal and judicial work in many forms, including legal research, translation, transcription, document review, case management and administrative support. However, in courts, the use of AI has the power to affect rights, liberty, privacy, access to justice and public confidence in the justice delivery system. Any framework for AI in courts must therefore balance technological advancement with judicial independence, accountability and fairness.
On June 3, 2026, the Supreme Court of India issued a notice inviting views and suggestions from stakeholders and the general public on the draft Regulations for Use of Artificial Intelligence in Courts, 2026 (draft regulations). The draft has been prepared under the aegis of the Artificial Intelligence Committee of the Supreme Court of India. Comments and suggestions have been invited by June 20, 2026, to the Member Secretary, AI Committee, Supreme Court of India. The draft regulations are intended to create an institutional framework for responsible AI adoption across India’s judicial system.
The draft regulations identify where AI may be used, where it must not be used, who must approve such use, what safeguards must apply, how AI-generated material must be disclosed and what remedies may be available if prohibited use causes harm. The draft, therefore, treats AI as a regulated judicial and administrative tool, not as an uncontrolled technological add-on.
Applicability and Commencement
The draft regulations are stated to be an addition to the Information Technology Act, 2000 and the Digital Personal Data Protection Act, 2023. Where another applicable law prevails on a subject, such law will continue to apply. The draft regulations apply to the Supreme Court of India, the High Courts, and all courts, including tribunals and statutory commissions, performing adjudicatory functions in India.
The commencement of the regulations is proposed to be staggered. For the Supreme Court, the regulations would come into force on the date notified in the Official Gazette by the Chief Justice of India. For High Courts and all courts, tribunals and statutory commissions under its jurisdiction, the regulations would come into force on the date notified by the Chief Justice of the respective High Court.
Further, different dates may be appointed for different provisions. This phased approach would give courts flexibility in implementation, while creating a common national framework on the use of AI in courts.
Key Definitions
The draft regulations lay down a detailed set of definitions, clarifying the scope of various terms, as the use of AI in courts may involve multiple technologies and associated risks.
The term “Artificial Intelligence” has been defined as a machine-based system that can infer, learn and generate decisions, predictions or recommendations from data with varying degrees of autonomy. General-purpose software is excluded unless it is specifically embedded with or dependent upon AI.
“AI System” and “AI Tool” have been defined to cover software, platforms, applications, devices or processes that use AI in connection with court processes.
The definition of “adjudicatory function” includes the application of law to facts for deciding rights, liabilities, status or obligations of parties.
“Administrative functions” include filing, registration, listing, notices, record maintenance, defect scrutiny and other registry functions.
The draft regulations define “Generative Artificial Intelligence” as AI that generates new content, including text, images, audio, video and code. It also requires disclosure and verification of generative AI content filed or produced before a court. This definition directly deals with the risk of fabricated citations, inaccurate legal summaries and synthetic material being placed before courts.
The definition of “hallucination” includes AI outputs that appear plausible but are factually incorrect, fabricated, misleading or unsupported by verifiable material. This is vital as an AI tool may produce confident but incorrect legal facts, evidence, case precedents, statutory provisions, rules, or legal principles.
The draft regulations define “Black Box” systems as AI systems whose internal processes and decision-making logic are not transparent or capable of being explained by identifiable rules or factors.
“Human-in-the-Loop” requires mandatory human review, supervision and verification of AI outputs, with final responsibility resting with a human.
The draft regulations define “Risk Scoring” as the use of AI to assign a score to an individual to estimate future behaviour, including recidivism, failure to appear before court or commission of an offence.
Judicial Independence and Human Accountability
One of the strongest themes of the draft regulations is that AI must remain compliant with human judgment. The draft regulations provide that AI systems would function only in an assistive capacity and not replace or compromise the independent exercise of judicial authority by a duly appointed judicial officer. The final authority to determine matters of law, fact and justice must remain with the competent judicial officer.
Further, any decision made with the assistance of AI would remain the responsibility of the concerned officer. The draft regulations expressly state that officers cannot rely on the output of an AI system, the opacity of a black box system or hallucination as grounds for avoiding accountability for an incorrect, illegal or harmful decision. AI may assist in research, summarisation or administrative processing, but it cannot carry judicial responsibility. The draft regulations place the burden of verification on the human user and protect the constitutional role of a judge.
Rule of Law, Fairness and Non-Discrimination
The draft regulations require the use of AI in courts to remain consistent with the Constitution, existing law, and principles of natural justice. AI cannot be used in a manner that undermines due process, fair trial rights, equality before law or access to justice. The draft also clarifies that the Bangalore Principles of Judicial Conduct will continue to govern judicial officers in matters involving AI.
The draft regulations further state that AI systems must be designed, trained, and deployed in a manner that avoids discrimination on the grounds of race, religion, caste, sex, gender, disability, language, and economic status. It also requires special care for vulnerable groups, including women, children, persons with disabilities, minority communities and persons from socially or economically disadvantaged backgrounds. This is an important provision as AI systems heavily depend on data. If the data is biased, incomplete or unrepresentative, the tool may reproduce or amplify unfair outcomes.
Transparency, Explainability and Verification
Under the draft regulations, AI systems used in court processes must meet high standards of transparency and explainability. Their functioning, data inputs, and decision logic must be understandable and easily explainable to judicial officers, parties and the public. Opaque systems would be subject to increased scrutiny, especially in cases where they affect personal liberty or rights. The draft also states that AI-generated output or information used in courts must be treated as advisory, and reasonable care must be taken to verify its accuracy before use. The framework places responsibility on the human user to ensure that court processes are not affected by fabricated, inaccurate or misleading AI-generated material.
Permissible Uses of AI in Courts
The draft regulations provide a detailed list of permitted uses of AI, subject to prior approval by the Appropriate Authority in writing and supervision by nominated officers.
- AI may be used for case management, including identification of defects in new filings, cause-list preparation, hearing scheduling and docket prioritisation. It may also be used for automated transcription of court proceedings, provided the accuracy is reviewed and certified by a Designated Officer.
- The draft permits AI-assisted translation of judgments, orders, pleadings, and other legal documents, but requires human verification. AI may also be used for legal research, precedent retrieval, citation verification and document summarisation. These are areas where AI can reduce time and improve access to information, provided the results are checked carefully.
- The draft regulations further allow the use of AI in administrative functions such as case-filing assistance, defect scrutiny, record management and judicial resource allocation. Conversational AI assistants and guided chatbots may be used to help litigants and stakeholders understand court services and procedural requirements, subject to human oversight. AI may also be used for text-to-speech, speech-to-text, Braille translation and visual assistance tools for persons with disabilities or language barriers.
- Other permitted uses include document authenticity verification, fraud detection in administrative processes, anonymisation of judgments and court records, court performance assessment, backlog monitoring and auto-generation of prescribed formats, notices and summons.
Prohibited Use of AI Systems
The draft regulations also lay out absolute prohibitions, which cannot be relaxed or modified by any authority.
- The draft prohibits any judgment, order or finding of fact or law from being reached through algorithmic decision-making alone or solely on the basis of AI-generated information, data or analysis. Human judicial authority must remain the determinative authority in all adjudicatory decisions.
- AI systems cannot perform adjudication or sentencing without mandatory human-in-the-loop safeguards. Any AI output relating to adjudicatory or sentencing questions must be advisory only and subject to independent judicial evaluation.
- The draft regulations also prohibit risk scoring in court processes. AI cannot be used to assess flight risk, predict recidivism, evaluate bail eligibility or determine the credibility of parties or witnesses.
- The use of undisclosed, opaque or unexplainable AI systems is prohibited where they may materially affect lawful rights or personal liberty.
- AI cannot be used to predict, profile or infer the future conduct of parties, accused persons, witnesses or legal representatives.
- AI-based surveillance or continuous monitoring of judicial officers, advocates, litigants or other persons connected with court premises or court processes is also prohibited, except in cases when such action is authorised.
- The draft regulations also provide that AI-generated output cannot be submitted as an independent source of evidence without full and transparent disclosure of its AI-generated character.
- Any use that compromises the confidentiality of judicial deliberations or judicial independence is prohibited.
In case of any violation, it must be reported to the AI Secretariat and placed before the AI Committee, which may order remedial measures, including suspension of the relevant AI system.
Disclosure of AI-Assisted Material
Where an AI tool assists in case management, document analysis or judicial administration in a manner that may affect proceedings, all parties involved must be informed in a timely and accessible manner. The draft regulations also create an obligation for parties and lawyers when an AI tool is used in preparing or submitting any document, pleading or evidence. In such cases, the AI-assisted character of that material must be disclosed to the court through a declaration or certificate. Court-initiated use of AI must also be declared in the prescribed format.
The court may also require disclosure of the AI system used, the nature and extent of AI assistance and the steps taken to verify the accuracy of AI-generated content. If any document, pleading or evidence is fabricated, false, misleading or inaccurate because of its AI-generated character, the person submitting it would be held responsible. Further, under the draft regulations, the AI-generated character of the output cannot be used as a defence.
Establishment of Apex Body for AI Governance
The draft lays down provisions for a permanent, full-time Apex Body at the Supreme Court of India. The Apex Body would be empowered to regulate and promote innovation, integration, governance, and oversight, set standards and develop policy in AI. The functions of the Apex Body would include ensuring that AI systems comply with the Constitution and law, maintaining a harmonised national approach, coordinating with High Court AI Committees, liaising with MeitY, NIC, CERT-In, the Data Protection Board and other bodies, considering grievances and publishing an annual governance report on the state of AI in Indian courts. Further, the Apex Body would be empowered to set minimum mandatory standards, approve AI systems proposed for use in court processes, issue guidelines and oversee the Centre of Research and Excellence on Artificial Intelligence.
The draft regulations also provide for committees of the Apex Body, including the Judicial Committee, Technical Committee, Committee on Infrastructure and Finance, Case and Data Management Committee and Cyber Security Committee. Each committee would have a specific function connected with judicial use, technical assessment, infrastructure, data management and cybersecurity.
The draft also establishes the Centre of Research and Excellence on Artificial Intelligence (CoRE-AI), which would conduct research, evaluate AI tools, maintain records of available tools and technical regulations, provide support to the Apex Body, track legal developments and organise academic and institutional discussions. Further, AI Committees and AI Secretariats at the Supreme Court and High Court levels would regulate AI adoption within their jurisdictions, approve AI systems after technical and ethical assessments, direct audits, review AI incidents and prepare annual reports.
Approval, Testing, Audits and Incident Management
The draft regulations provide operational safeguards for AI systems used in courts, including a comprehensive Technical and Ethical Impact Assessment prior to approval of any AI system. Controlled environment testing may also be permitted before full deployment of the AI system. Each court would also be required to maintain an AI Register, which would document approved AI systems, their purposes, service providers, approval dates, conditions, impact assessments, audits and AI incidents. The draft regulations also provide for periodic technical, legal and ethical audits at intervals not exceeding 1 year, or at shorter intervals directed by the Appropriate authority. Every AI Secretariat would also be required to maintain an AI Incident Database recording the type, cause, consequences and remedial measures relating to AI incidents.
Private Sector Engagement and Court Data
The draft regulations state that no private vendor or third-party service provider may participate in an AI system deployed in court processes without prior written approval of the Appropriate Authority. Private sector proposals must be evaluated for technical capability, legal compliance, ethical standards, data security practices and financial standing. Contracts with private entities must include safeguards on ownership and access to court data, restrictions on use of sensitive judicial data, incident reporting, audit, consequences of breach, source and model transparency, explainability, indemnity and deployment requirements for systems processing sensitive judicial data.
Additionally, when AI tools are developed using court data or court resources, the Appropriate authority must ensure that the court retains ownership or a perpetual royalty-free licence to the tool and outputs. Private entities cannot claim exclusive IP rights over tools developed using judicial data or public resources.
Data Protection and Cybersecurity
The draft regulations require AI systems used in court processes to comply with the Digital Personal Data Protection Act, 2023, the Information Technology Act, 2000 and other applicable laws. Further, sensitive judicial data cannot be transferred to any external system without the written authorisation of the Appropriate Authority. Personal data would need to be anonymised to the extent technically feasible before it is used for training, testing or refinement of AI systems. The draft also requires regular cybersecurity audits at intervals of one year and an annual review of access protocols.
Training, Best Practices and Institutional Capacity
Under the draft regulations, regular structured training for judges, advocates and court staff who are required to use or interact with AI systems would be required. Training would cover the functioning and limitations of AI systems, bias, hallucinations, technical errors, legal and ethical obligations, data protection, cybersecurity and incident reporting.
Grievance Redressal
The draft regulations state that when some harm is caused to a party as a direct or indirect effect of prohibited AI use under Regulation 20, the party or legal representative may file an application before the court in which the AI system was or is being used. The court must give a reasonable opportunity of hearing and may pass appropriate orders. High Courts may also lay down principles and procedures for dealing with grievances arising from AI use in court processes.
Conclusion
The draft Regulations for Use of Artificial Intelligence in Courts, 2026 is a critical initiative by the Supreme Court towards formalising AI governance in the judicial system. The draft regulations recognise that AI is a useful tool to support court administration, legal research, translation, transcription, accessibility and information management. However, the draft regulations place firm limitations on the use of in decision-making, sentencing, risk scoring, profiling, surveillance and rights-affecting decisions.
Authors: Manisha Singh and Shivi Gupta