The Supreme Court’s AI Committee has released a comprehensive set of draft rules titled the Regulations for Use of Artificial Intelligence in Courts, 2026. The proposed framework intends to regulate the deployment of Artificial Intelligence (AI) systems across the country’s entire judicial spectrum, encompassing the Supreme Court, High Courts, subordinate courts, tribunals, and statutory commissions performing adjudicatory functions.
While the guidelines actively encourage the integration of technological tools to optimize efficiency and minimize systemic delays, they institute absolute boundaries to ensure human primacy in core judicial functions. The draft text has been placed in the public domain, with stakeholders and the general public invited to submit comments and suggestions until June 20.
Permitted Systems and Strict Operational Boundaries
The draft framework outlines a specific list of judicial and administrative functions that may be assisted by AI, provided they operate under constant human supervision and receive appropriate official approval. Permitted operations include legal research, precedent retrieval, citation verification, and the automated transcription of court proceedings. Furthermore, the rules allow AI to be used for case management, defect scrutiny, cause list preparation, record management, translation of orders, and generating prescribed formats for notices and summons. The regulations also encourage the development of accessibility tools for persons with disabilities and AI-powered chatbots to help litigants navigate court services.
Conversely, the regulations draw non-negotiable lines where the use of AI is completely barred to protect independent judicial authority. AI systems are strictly prohibited from deciding cases, passing sentences, determining bail eligibility, or evaluating witness credibility. The framework specifically bars courts from utilizing AI-based risk-scoring models to assess flight risks, the likelihood of reoffending, or whether an accused person might commit future offenses. Furthermore, AI cannot be used to conduct surveillance on judges, advocates, or litigants, nor can it interfere with judicial deliberations or employ opaque, unreviewable decision-making architectures in matters affecting rights or personal liberty.
Registration Begins for FutureCrime Summit 2026, India’s Largest Cybercrime Conference
Mandatory Disclosure Requirements and Full Litigant Accountability
Under Regulation 43(3), a strict disclosure mandate is introduced for all active legal proceedings. If a party or a legal representative employs any AI system to prepare or submit documents, pleadings, or evidence, the AI-assisted nature of the material must be declared to the court at the exact time of submission. Under Regulation 43(4), judges are fully empowered to demand specific technical details regarding the AI platform used, the depth of assistance received, and the exact steps taken by the counsel to verify the baseline accuracy of the outputs. The guidelines make it clear that AI-generated information cannot be treated as an independent source of evidence on its own.
The framework firmly establishes that human litigants and lawyers retain complete professional accountability. Regulation 43(6) states that if any pleading or filing is found to be fabricated, false, misleading, or inaccurate due to its AI-generated character, the individual who submitted it will bear full responsibility. Courts will bar individuals from using the automated nature of an AI platform as a valid defense. Additionally, the text explicitly recognizes the phenomenon of AI hallucinations, where systems output plausible-sounding but completely fabricated legal precedents, statutory provisions, or facts. Court officials and legal officers are barred from using system opacity or algorithmic errors as a shield against incorrect, illegal, or harmful decisions.
New Governance Structure and Strict Private Vendor Rules
To maintain uniform oversight, the regulations propose a new governance architecture, starting with a permanent apex body at the Supreme Court level. This primary council will comprise Supreme Court judges, High Court Chief Justices and judges, technology experts, cybersecurity specialists, finance professionals, and advocates specializing in technology laws. It will be responsible for setting national standards, evaluating software, and publishing annual reports. Additionally, dedicated AI Committees and Secretariats will be established across every High Court, alongside a specialized Center of Research and Excellence on Artificial Intelligence (CoRE-AI) to provide technical support. To oversee generative content protocols, a dedicated AI Content Verification Authority will also be formed.
Every deployed AI system must undergo comprehensive technical, legal, and ethical audits at intervals not exceeding one year. Courts must maintain a live AI Register and an active AI Incident Database to log all system errors, biases, and security breaches, with serious errors requiring immediate reporting. Finally, severe restrictions are placed on private technology companies. No private vendor can participate without prior approval, and all contracts must guarantee that the judiciary retains full ownership of court data, strict audit rights, and clear liability provisions for AI-related harm. If tools are trained using court resources, the courts must retain permanent ownership or a perpetual royalty-free license to the tools and their outputs. All systems must strictly comply with the Digital Personal Data Protection Act, 2023, barring the use of personal records for AI training without explicit, authorized clearance.
