Supreme Court of India, draft Regulations for Use of Artificial Intelligence in Courts, 2026
A draft instrument published for public consultation on 3 June 2026 by the Supreme Court's Artificial Intelligence Committee, grounded in human primacy, transparency, accountability, data protection and judicial independence, that would make an officer exclusively accountable for AI-assisted decisions, treat AI output as advisory subject to verification, prohibit algorithmic adjudication and risk scoring, and require parties to disclose AI-assisted filings at submission. A draft as of 25 September 2026, not law.
When this decides something for you
An advocate preparing a filing with machine assistance, a judge deciding how far a court tool may be used, a registrar procuring a system, or a reader abroad comparing India's approach with the EU's: each needs to know what the draft would require if adopted, what it already signals, and what it is not. As of the date of this page it is a draft. Its consultation window closed on 20 June 2026. No final instrument has been located.
What the instrument is
A notice of the Supreme Court of India dated 3 June 2026, issued by the Member Secretary of the AI Committee, published a preliminary draft of "Regulations for Use of Artificial Intelligence (AI) in Courts, 2026" and invited comments by 20 June 2026. The notice states the draft's grounding: "the principles of human primacy, transparency, accountability, data protection, and judicial independence". The draft runs to five chapters: preliminary; general principles; permissible and prohibited uses; policy-making and institutional mechanism (a permanent Apex Body at the Supreme Court, High Court AI Committees and secretariats); and operational provisions including transparency, disclosure and an AI Content Verification Authority.
What it provides
Human primacy (Regulation 4). AI in court processes "shall at all times remain strictly subservient to human judgment and judicial authority"; every system functions "solely in an assistive capacity"; the ultimate authority on law, fact and justice "shall vest exclusively in the judicial officers".
Accountability (Regulation 8). Accountability for all decisions made by any officer with AI assistance "shall rest exclusively upon such officer", and it is not permissible to invoke the system's output, the opacity of a black-box system or the occurrence of hallucination "as a ground for avoiding accountability for a palpably incorrect, illegal, or harmful decision". Under Regulation 8(3), AI-generated output used in any court "shall be treated as advisory in nature and reasonable care shall be taken to verify the accuracy of such output before the same is utilised". Two provisos follow: the responsible officer "may, for reasons to be recorded in writing, dispense with the requirement of verification"; and tools used exclusively for administrative, non-adjudicatory functions and certified as reliable by the AI Secretariat are deemed verified on a class basis.
Prohibited uses (Regulation 20). Stated as "absolute and non-derogable": no judicial outcome reached "through Algorithmic Decision-Making alone or solely on the basis of AI-generated information"; no adjudication or sentencing by a system "without mandatory Human-in-the-Loop"; no risk scoring for flight risk, recidivism, bail eligibility or "the credibility of parties or witnesses"; no undisclosed or unexplainable system in any process that may materially affect rights or liberty; no prediction or profiling of parties, accused, witnesses or representatives; no surveillance of judges, advocates or litigants except as law authorises; no AI-generated output submitted "as an independent source of evidence without full and transparent disclosure of its AI-generated character"; nothing that compromises the confidentiality of deliberations or the independence of decision-making. Violations are reported to the AI Secretariat and dealt with by the AI Committee (Regulation 21).
Disclosure (Regulation 43). Courts must inform parties where a tool materially assists case management, document analysis or administration. Under Regulation 43(3), where a tool is used by any party or representative "in the preparation or submission of any document, pleading, or evidence, the AI-assisted character of such material shall be disclosed to the Court at the time of submission" by a declaration or certificate in a prescribed form. Under 43(4) the court may require disclosure of the system used, the nature and extent of assistance, and "the steps taken to verify the accuracy of any AI-generated content". Under 43(6), where AI-generated material is found fabricated, false, misleading or inaccurate, "the person submitting the same shall bear full responsibility therefor and shall not be entitled to rely upon the character of the AI output as a defence".
What it signals and what it is not
The draft states, as a proposed rule of the court system, the two propositions this library's canon carries: that responsibility rests on the officer and never on the tool, and that verification precedes reliance. It also introduces the first proposed filing-side disclosure duty for Indian advocates, which, read with the Bar Council's publication-side disclosure paragraph of July 2026, means disclosure is arriving from both directions.
Two things it is not. It is not law: nothing in it binds until notified after due process, and the Supreme Court itself said so in Gadiya on 2 September 2026, noting the regulations were "yet to be finalised". And Regulation 8(3)'s first proviso is a verification waiver, not a waiver of written reasons: an officer may dispense with verification for reasons recorded in writing. The critique of that clause, where made, is of the clause and not of the Court.
Where it stands in the record
The draft is a dated instrument. Its status is checked before any page cites it as more than a draft, and this page will be superseded, not edited, when a final instrument appears.
Provenance and method.
Where this was published
- Supreme Court of India
- Notice dated 03.06.2026 and draft Regulations for Use of Artificial Intelligence (AI) in Courts
- 2026 (archived copy A3
- read 25 September 2026)
Changelog
- 2026-09-25, version 1. First published.