Supreme Court extends consultation on AI rules for courts: What the draft proposes

The Supreme Court has extended the time till July 15, 2026, for the submission of comments by stakeholders and the general public on the draft ‘Regulations for Use of AI in Courts’. A preliminary draft was issued on June 3 by the AI Committee of the Supreme Court.

As per the Union Minister for Law and Justice Arjun Ram Meghwal's reply in the Lok Sabha on December 5, 2025, “an amount of ₹53.57 crore has been earmarked for the component Future Technological Advancements (AI, Blockchain, etc.)” under the e-Courts Project Phase-III, to integrate modern technologies for a smoother user experience.

The genesis of AI in judiciary dates back to 2005, when the eCourt Project was started to create a digital backbone through CIS, NIC and the NJDG. However, this infrastructure was designed for the digitisation of documents and did not contemplate the use of AI tools in the judicial system.

We have thereafter seen an ad-hoc and unsystematic penetration of AI in the judicial system for the translation of documents and other miscellaneous purposes: Supreme Court Vidhik Anuvaad Software ('SUVAS') was developed with technical support from MeitY; AI 'Saransh', which is used for the summarisation of case records, was developed by NIC, a private company ManCorp Developed 'SUPACE', an AI-enabled tool to help the judges to examine the large volumes of case records and 'Adalat' was created by a private vendor for AI based transcription and translation services, which is used by courts across 9 states.

The draft regulation states that the adoption and use of AI in courts shall be consistent with the provisions of the Constitution and the laws and principles of natural justice. It shall not be used in a manner that undermines due process, the right to a fair trial, equality before the law, or access to justice.

The advancement of AI at an unprecedented and exponential rate is posing a dual threat to countries and democratic governance. The government is setting up a unified centre of command to address cybersecurity issues arising from advancements in AI and general-purpose AI models such as Anthropic's Claude Mythos. As such, it is essential to contextualise the use of AI in courts through such vague guidelines:

Constitutional mandate: As per Article 145(1) of the Constitution, “subject to the provisions of any law made by Parliament, the Supreme Court may from time to time, with the approval of the President, may make rules". To this end, the Supreme Court Rules, 2013, were made to govern the practice, procedure, and administration of the Apex Court. However, unlike the 2013 Rules, the draft regulations do not cite any constitutional authority and their enforceability is, as such, legally suspect.

This is borne out by Chief Justice Surya Kant's recent statement, in which he said that the Supreme Court is not a headmaster and that the High Court's independent constitutional space must be respected. It is perhaps due to this reason that the Supreme Court is not passing a judgment under Article 142 of the Constitution to establish an All-India Judicial Service (AIJS), which could enhance administrative efficiency and facilitate judicial reform. However, the draft regulations for AI do not reflect this constitutional mandate. These rules will be notified by the respective Chief Justice of each High Court and will be applicable to all district courts, tribunals and statutory commissions. 

As per CJI Suryakant, High Courts have vast powers under Articles 226 and 227 and that they may wield broader jurisdictions. So these rules should govern the use of AI in the Supreme Court alone and not extend to High Courts and District Courts.

Data Security: The Bar Council of India has stated that under the Advocates Act, 1961, AI lacks legal recognition and thus, only lawyers can be held responsible for any erroneous content generated by AI. Due to mounting concerns, in April 2026 the Punjab and Haryana High Court issued a circular barring all judicial officers from using AI tools, which include ChatGPT, Gemini and Copilot for legal research or writing of judgments.

Presently, the draft regulation uses the term sensitive judicial data, which may include personal identifiable information of parties, witnesses or legal representatives and any information processed in connection with a Court process, the unauthorised disclosure of which may cause harm. It goes on to state that “AI systems used in Court processes shall be trained and operated based on data that is accurate, representative, lawfully obtained and to the extent feasible, free from discriminatory bias”. It further says that the “deployment of AI Systems, trained on unlawfully collected or demonstrably biased datasets, shall be prohibited”. 

The P&H High Court emphasised that AI poses critical risks such as generating fabricated case laws or hallucinations. The rising privacy concerns emanating from the use of AI should be read in the context of the Constitution Bench judgment in the Puttaswamy matter, which declared the right to privacy to be a fundamental right.

However, even to this date, the Digital Personal Data Protection Act (DPDP Act), which was passed in 2023, is yet to come into force. The P&H High Court emphasised this point and held that feeding sensitive case-related information into public AI models violates data privacy and confidentiality. The sensitive judicial data of millions of litigants cannot be given to private AI players without enforcing strict data protection law in India.

Capacity and Remit: Considering fundamental capacity constraints in district courts, how feasible is it to expect an enabling infrastructure for appropriate AI usage to emerge overnight? Driven by the exigencies of the COVID-19 pandemic, Indian courts rapidly, and often in an ad hoc manner, adopted newer technologies and began offering online hearings and live streaming. The absence of clear regulations on data use and security, especially depending on foreign platforms like Zoom, WhatsApp and YouTube, is in violation of the Supreme Court judgment of 2018 in Swapnil Tripathi and CASC matter.

In the Govindacharya matter, former Chief Justice of India D.Y. Chandrachud stated in November 2022 that the Supreme Court will shortly develop its own tech platform to protect institutional integrity and prevent the circulation of distorted video clips on social media, a promise which remains unfulfilled.

Prime Minister Modi, while talking about the need for “ease of justice”, has called for the use of AI & technology to facilitate the release of impoverished undertrials and alleviate Indian courts’ significant backlog. However, a roadmap for judicial reform to unlock this is missing in the Apex Court’s draft regulations on AI.

 (The author is an advocate, Supreme Court of India)

The opinions expressed in this article are those of the author and do not purport to reflect the opinions or views of THE WEEK.