India’s judiciary is preparing for what could become its most significant technological transformation, but with a clear constitutional safeguard: artificial intelligence will remain an assistant, never a judge. The Draft Regulations for Use of Artificial Intelligence in Courts, 2026, released by the Supreme Court’s AI Committee for public consultation, seek to establish a uniform framework governing AI across the Supreme Court, High Courts, subordinate courts, tribunals and statutory adjudicatory bodies.
The proposed regulations acknowledge AI’s growing role in legal practice while making one principle non-negotiable—the power to decide questions of law, fact and justice will continue to rest exclusively with human judges.
The draft permits lawyers and litigants to use AI tools for legal work, but introduces a mandatory disclosure regime. Any party relying on AI to prepare pleadings, documents, submissions or evidence must disclose its use at the time of filing. Courts may also seek details of the AI platform used, the extent of its assistance and the steps taken to verify the accuracy of its output. Responsibility for errors or AI “hallucinations” cannot be shifted to technology; accountability remains firmly with legal professionals and judges.
The proposed framework reflects a balanced approach—encouraging innovation while guarding against the risks posed by increasingly sophisticated generative AI systems. It identifies five guiding principles: primacy of human decision-making, transparency, accountability, data protection and judicial independence.
Under the regulations, AI may be used for legal research, citation verification, drafting assistance, translation, transcription, scheduling, case management, record maintenance and judicial administration. Courts are also encouraged to deploy AI-powered accessibility tools and litigant-facing services that simplify navigation of the justice system and improve access for persons with disabilities.
Equally significant is the detailed list of activities AI will never be allowed to perform. The draft absolutely prohibits AI from deciding cases, recommending bail, determining sentences, evaluating witness credibility, predicting the future conduct of litigants or influencing judicial deliberations. It also bars courts from relying on opaque “black-box” algorithms in matters involving personal liberty or fundamental rights, including AI-based risk-scoring systems increasingly used elsewhere in legal practice.
The emphasis on human control follows growing concern over AI-generated misinformation entering court proceedings. The Supreme Court’s move comes after several incidents where fabricated case citations generated by AI found their way into legal pleadings and even judicial orders. In one widely discussed instance, a trial court relied on non-existent judgments allegedly produced through artificial intelligence, prompting the Supreme Court to caution against the uncritical use of AI in judicial processes.
Similar concerns surfaced earlier in the Delhi High Court. In 2023, the Court refused to rely on arguments generated through ChatGPT in a trademark dispute, observing that large language models were capable of fabricating authorities and factual material. In another matter, homebuyers withdrew a petition after discovering that portions of their pleadings—including fictitious case citations—had been generated through AI without proper verification. The episodes reinforced the judiciary’s growing consensus that while AI offers speed and convenience, it cannot replace legal diligence.
Reflecting that philosophy, Regulation 4 declares that AI “shall at all times remain strictly subservient to human judgement and judicial authority”. Every AI system, it says, must function solely in an assistive capacity, while the authority to determine questions of law, fact and justice “shall vest exclusively in the judicial officers of the competent jurisdiction”.
Implementation of the framework would be overseen through an elaborate institutional structure. The draft proposes a permanent apex AI body at the Supreme Court to regulate innovation, standard-setting, governance and oversight, supported by AI Committees and dedicated AI Secretariats across High Courts. Annual audits of AI systems, cybersecurity safeguards, AI registers, incident databases and compliance with data protection laws would become mandatory, alongside regular training programmes for judges, lawyers and court staff.
The committee behind the proposal—headed by Justice PS Narasimha and comprising Justices Sanjeev Sachdeva, Raja Vijayraghavan V Anoop Chitkara and Suraj Govindaraj—has invited comments from stakeholders before the regulations are finalised.
The timing is significant. India’s courts currently face a backlog of nearly 52 million pending cases, making technology an increasingly attractive solution for improving efficiency. AI-assisted transcription, translation, document management and scheduling could substantially reduce administrative burdens, allowing judges to devote more time to adjudication.
Yet, the legal profession remains cautious. A recent Manupatra Academy survey found that more than half of legal professionals already use AI tools, while nearly three-fourths have experimented with generative AI in their work. At the same time, respondents cited serious concerns over data privacy, client confidentiality and the reliability of AI-generated content—precisely the issues the draft regulations seek to address.
The proposed framework, therefore, represents neither blind enthusiasm nor technological scepticism. Instead, it acknowledges that AI is becoming an indispensable part of modern legal practice while insisting that constitutional values cannot be delegated to algorithms.
For a judiciary long associated with paper-heavy procedures and mounting delays, AI promises faster administration, improved accessibility and greater efficiency. But the draft sends an equally powerful message: justice is ultimately a human responsibility. Technology may assist the process, but it cannot replace judicial conscience.
—The writer is former Senior Managing Editor, India Legal magazine
The post Justice, Assisted appeared first on India Legal.