The Supreme Court’s proposed “Regulations for the use of Artificial Intelligence (AI) in Courts, 2026” could be one of the most consequential judicial reform initiatives in recent years. The draft framework, released during the tenure of Chief Justice of India (CJI) Surya Kant, is not just a technology policy but also an attempt to answer the question that courts around the world are grappling with: How should the justice system harness the benefits of AI without compromising judicial independence, due process, impartiality and human judgment?
What makes the draft particularly noteworthy is that it neither treats AI as a threat to be countered nor as a miracle solution to judicial delays. Instead, it takes a carefully balanced approach of encouraging innovation while putting clear red lines around areas where machines cannot be allowed to intrude. The rules apply not only to the Supreme Court and High Courts but also to district courts, tribunals and statutory adjudication bodies across the country.
The proposed framework is significant as India is one of the first major jurisdictions to attempt to create a comprehensive judicial AI governance architecture before large-scale deployment of AI tools in courts becomes common.
AI can help, not judge
The single most important principle running through the draft regulations is what the Supreme Court calls “humanitarian primacy and judicial independence.” The rules clearly declare that AI must remain “strictly subject” to human judgment and judicial authority. The power to determine questions of law, facts and justice will remain exclusively vested in the judges.
This is important as courts around the world are increasingly experimenting with AI-assisted decision making. In some jurisdictions, algorithmic tools have been used to assess bail risks, sentencing patterns, and probabilities of recidivism. Such systems have generated intense controversy due to concerns about hidden biases and lack of transparency.
The Supreme Court’s draft rules strongly reject this path. The proposed rules prevent AI from deciding cases, sentencing or delivering judicial outcomes through algorithmic decision making. Even where AI assists, its outputs should remain advisory and subject to independent judicial scrutiny. Judges cannot shirk their responsibilities by relying on machine-generated recommendations.
The framework also recognizes concerns regarding “black-box” systems – AI models whose internal logic cannot be explained. It prohibits the use of opaque or unclear AI systems in matters that affect legal rights or individual freedoms. In fact, the Supreme Court is drawing a constitutional limit that AI can help judges work faster, but it can never become a judge.
What AI can and cannot do inside the courts
The draft rules allow for a wide range of AI-assisted actions designed to improve efficiency and access to justice. These include legal research, citation verification, summarizing arguments and judgments, translation, transcription of court proceedings, drafting assistance, hearing scheduling, records management, and case administration. AI-powered chatbots can also be used to help litigants understand procedures and access court services. Accessibility devices are especially encouraged for individuals with disabilities.
The proposed rules also create “a presumption in favor of responsible AI adoption”, indicating that courts should actively explore technologies capable of reducing delays and improving judicial administration. The draft further states that where appropriate safeguards are in place, innovation should generally be prioritized over restraint.
Also, the list of prohibited uses is shocking. AI cannot be used to assess flight risk, predict recidivism, determine bail eligibility, or evaluate the credibility of witnesses and parties. Courts are also barred from using AI for profiling litigants, predicting future behavior or monitoring judges, lawyers and litigants. The rules prevent AI-generated content from being considered as independent evidence unless its AI-generated nature is fully disclosed.
These restrictions reveal the Supreme Court’s awareness of the dangers posed by predictive algorithms. The draft effectively rejects the idea that machines should assess human character, future conduct or criminal propensity.
Disclosure, accountability, and the problem of AI hallucinations
Perhaps the most practical feature of the draft rules is the accountability framework governing AI-assisted filing.
Lawyers and litigants will be allowed to use AI to prepare pleadings, documents and submissions. However, if AI has been used, that fact must be disclosed to the court through a prescribed declaration. Courts may also require parties to disclose the specific AI system used, the extent of AI assistance, and the steps taken to verify the accuracy of AI-generated content. This is a direct response to the growing global problem of lawyers citing non-existent decisions generated by AI systems.
The rules recognize the risk of fabricated, inaccurate or misleading AI output. If any pleading, document or evidence is found to be false due to AI-generated content, the responsibility will entirely rest on the person filing it. The plaintiff or lawyer cannot avoid liability by claiming that the AI system made a mistake.
The framework also considers the disclosure of synthetic or AI-generated information used during proceedings and proposes an AI content verification authority to develop standards and protocols to verify generative AI outputs.
In short, the message is clear: AI can assist in advocacy, but the responsibility remains human.
A new governance architecture for judicial AI
The most forward-looking aspect of the draft may be the detailed institutional framework proposed to govern AI in courts.
The rules envisage a permanent apex body at the national level to set standards, approve AI systems and monitor AI adoption in the judiciary. This body will be supported by special committees dealing with judicial applications, technology, infrastructure, finance, cyber security and data management. A dedicated Center of Research and Excellence on Artificial Intelligence (CORE-AI) is also proposed to provide technical and legal support.
Each High Court will have its own AI Committee and AI Secretariat, headed by a judicial officer and assisted by technology experts. These bodies will monitor implementation, approve AI tools, investigate incidents and conduct periodic reviews.
Prior to deployment, AI systems will undergo technical and ethical impact assessments evaluating training data, risk of bias, hallucinations, cybersecurity vulnerabilities, and compliance with human oversight requirements. Courts may also require controlled environmental testing before large-scale implementation.
The draft regulations mandate AI audits, AI registers, incident reporting systems and annual transparency reports. High courts, tribunals and commissions will be required to publicly disclose the AI systems they use, audit results and incidents recorded.
The framework also imposes stringent safeguards on private sellers. No private entity can participate in the court AI system without prior approval. Contracts should include provisions on ownership of court data, restrictions on data use, audit rights, transparency obligations and liability for damages.
The Supreme Court’s draft AI rules represent an ambitious effort to prepare the Indian judiciary for the technological changes that are already reshaping legal systems around the world. Rather than wait for AI to be adopted in an unregulated manner, the top court has sought to establish constitutional guardrails before deeply incorporating the technology into judicial processes.
The proposed framework believes that AI can help courts tackle delays, improve access to justice and increase administrative efficiency. Yet it simultaneously emphasizes that adjudication remains a fundamentally human task. By combining the encouragement of innovation with strong safeguards on accountability, privacy, transparency and judicial independence, the draft rules offer a model that is both technically progressive and constitutionally cautious.
If implemented after public consultation, the rules could become the foundation of India’s judicial AI ecosystem, and perhaps, a reference point for courts around the world to deal with the promises and perils of artificial intelligence.
