Supreme Court Releases Draft AI Rules For Courts; Lawyers Must Disclose Use Of AI In Pleadings

Supreme Court Drafts AI Rules For Courts

The Supreme Court has placed draft "Regulations for Use of Artificial Intelligence (AI) in Courts, 2026" in the public domain for comments until June 20, 2026. The framework, prepared under the Supreme Court's AI Committee, sets out how AI may be used across courts and tribunals while asserting that technology assists the judiciary but does not replace judges.

A key proposal requires lawyers and litigants to disclose AI-assisted material at the time of filing. Under draft Regulation 43(3), parties using AI to prepare pleadings, documents, submissions or evidence must flag this, and courts may seek details of the system used, the extent of assistance and verification steps taken. The draft also states that a party cannot escape liability by blaming AI; whoever submits false, fabricated or misleading content bears full responsibility.

The draft is built on "human primacy". AI must remain assistive and cannot decide cases, determine outcomes, pass sentences, assess witness credibility, decide bail or interfere with judicial deliberations. Judicial accountability stays with judges and court officials. AI "hallucinations" are treated as advisory and require human verification.

Permitted uses include legal research, citation verification, document summarisation, translation, transcription, drafting help, case management, cause-list preparation, scheduling, record management, accessibility services and litigant information tools. Draft Regulation 16 urges courts to adopt AI that improves access to justice and efficiency without replacing human decision-making or predicting dispute outcomes. The draft creates a presumption favouring responsible AI adoption.

Why It Matters For Businesses

The disclosure mandate will affect how companies, law firms and compliance teams prepare filings and evidence. Any AI-assisted drafting submitted to a court must be flagged, and the submitting party remains accountable for accuracy. This raises …