New Delhi, June 5 -- Seeking to strike a balance between technological innovation and judicial independence, the Supreme Court has proposed sweeping regulations governing Artificial Intelligence (AI) in courts, requiring lawyers to disclose AI-assisted filings and imposing an absolute ban on the use of AI for judicial decision-making, sentencing and bail determinations. The draft Regulations for Use of Artificial Intelligence in Courts, 2026, placed in the public domain for comments till June 20, state that AI systems may function only in an assistive capacity and cannot supplant the role of judges. The regulations underscore that "the ultimate authority to determine matters of law, fact and justice shall vest exclusively in the judicial officers of the competent jurisdiction." "These regulations aim to govern the use of AI, grounded in the principles of human primacy, transparency, accountability, data protection, and judicial independence, while establishing an institutional framework for responsible AI adoption across India's judicial system," the apex court said in its notice. The framework was prepared under the aegis of the Supreme Court's AI Committee and seeks to establish uniform standards for the use of AI in the Supreme Court, high courts, subordinate courts, tribunals and statutory commissions exercising adjudicatory functions. One of the most notable provisions is Regulation 20, which sets out a list of uses that are absolutely prohibited. The draft bars AI systems from independently determining judicial outcomes, passing sentences or performing adjudicatory functions. Any AI-generated recommendation relating to adjudication or sentencing would remain merely advisory and must undergo independent judicial evaluation. The regulations further prohibit AI-based risk scoring systems from being used to assess flight risk, predict recidivism, determine bail eligibility or evaluate the credibility of parties and witnesses. Courts are also barred from using AI to predict or profile the future behaviour of litigants, accused persons, witnesses or lawyers. The framework additionally prohibits the deployment of opaque or unexplainable "black box" AI systems in matters affecting legal rights or personal liberty, while also banning AI-based surveillance or continuous monitoring of judges, lawyers and litigants unless specifically authorised by law. At the same time, the draft adopts what it describes as a "presumption in favour of responsible AI adoption", encouraging courts to deploy AI tools that improve access to justice, reduce delays and enhance administrative efficiency. AI may be used for legal research, precedent retrieval, citation verification, document summarisation, translation of judgments and pleadings, transcription of proceedings, case management, scheduling, accessibility services and AI-powered litigant assistance tools, subject to human supervision and approval. Another key feature is Regulation 43 dealing with transparency and disclosure. The draft proposes that whenever a lawyer or litigant uses an AI tool in preparing pleadings, documents, submissions or evidence, the AI-assisted nature of the material must be disclosed to the court through a prescribed declaration. Courts would also be empowered to seek details regarding the AI system used, the extent of AI assistance and the verification steps undertaken before filing the material. The regulations clarify that litigants and lawyers cannot escape responsibility by blaming AI-generated errors. If any document, pleading or evidence is found to be fabricated, misleading or inaccurate because of its AI-generated character, the person submitting it will bear full responsibility and cannot rely on the AI-generated nature of the material as a defence. The draft expressly recognises the phenomenon of AI "hallucinations" and provides that accountability for decisions assisted by AI shall continue to rest with the concerned judicial officer. The regulations state that neither AI outputs nor system opacity can be invoked to avoid responsibility for an incorrect or unlawful decision. To oversee implementation, the framework proposes the creation of a permanent apex body at the Supreme Court level comprising judges, chief justices, technology experts, cybersecurity specialists, finance experts and advocates with expertise in technology law. It also envisages dedicated AI Committees and AI Secretariats in the Supreme Court and every high court, along with a Centre of Research and Excellence on Artificial Intelligence (CoRE-AI). The draft regulations have been published for stakeholder consultation, with comments and suggestions invited until June 20 before the framework is finalised....