Regulations for Use of Artificial Intelligence in Courts, 2026

The Supreme Court AI Committee, chaired by Justice P.S. Narasimha, has released the preliminary draft of the ‘Regulations for Use of Artificial Intelligence (AI) in Courts, 2026’. The committee has invited public and stakeholder comments on the draft related to use of Artificial Intelligence in Courts till June 20, 2026.

This move comes in the wake of growing concerns over judicial reliance on AI, highlighted by a recent Supreme Court observation where a trial court was chided for using non-existent, AI-generated judgments—an act the top court termed as “judicial misconduct” rather than a mere error.

Key Highlights of the Draft Regulations, 2026

Subservient to Human Judgment: The draft explicitly mandates that AI must function solely in an assistive capacity and remain strictly subservient to human judicial authority.
Prohibition on Core Judicial Outcomes: AI is strictly barred from determining sentencing, evaluating bail eligibility, assessing flight risk, or predicting recidivism (risk scoring).
Ban on Opaque Systems: The use of “black-box” or unexplainable AI systems is completely disallowed in any court process to ensure transparency.
Anti-Profiling and Surveillance: AI systems cannot be used to profile witnesses or parties, nor can they be deployed for continuous surveillance of judicial officers, advocates, or litigants unless specifically authorized by law.
‘Human-in-the-Loop’ Approach: Applications posing high risks to personal liberty or lawful rights will require mandatory human oversight and independent audits.
Data Protection & Non-Bias: Data processing must align with the Digital Personal Data Protection (DPDP) Act, 2023. AI systems must not amplify or introduce biases prohibited under the Constitution (caste, religion, sex, etc.).
Permitted Administrative Uses: AI can be deployed for non-decision-making tasks such as case management, transcription, translating judgments, and scheduling hearings.
Institutional Framework: The draft proposes a full-time Apex Body at the Supreme Court level, comprising Supreme Court and High Court judges, MeitY officials, and cybersecurity/finance experts to oversee policy development and standard-setting.

Need for Regulating AI in the Judiciary

Risk of “Hallucinations”: Generative AI is prone to fabricating data, citations, and precedents (as seen in the trial court incident), which compromises legal accuracy.
Algorithmic Bias: AI models trained on historical data run the risk of perpetuating systemic biases against marginalized communities, affecting fair trial principles.
Threat to Principles of Natural Justice: Decisions made by opaque, unexplainable algorithms violate a litigant’s right to a reasoned order and transparent justice.
Digital Divide: Over-reliance on highly technical tools could alienate rural, economically disadvantaged, or linguistically diverse communities, limiting access to justice.

Concerns and Challenges in Integrating AI in Judiciary

Risk of AI “Hallucinations”: Generative AI tools are notorious for fabricating facts, non-existent case laws, and legal precedents. If unchecked, relying on such synthetic data compromises factual accuracy and legal sanctity—a risk the Supreme Court recently classified as potential “judicial misconduct”.
Algorithmic Bias and Discrimination: AI systems are trained on historical data, which may contain deep-rooted social biases. Implementing such tools risks automating and amplifying prejudices related to race, religion, caste, or gender, directly violating the Right to Equality (Article 14).
The “Black Box” Dilemma (Lack of Transparency): Many advanced AI algorithms are opaque, meaning their internal decision-making paths cannot be audited or explained. Utilizing unexplainable AI in matters of personal liberty deprives litigants of a reasoned order, violating the fundamental principles of natural justice.
Widening the Digital Divide: While metropolitan courts might smoothly adopt cutting-edge tech, subordinate and rural courts often struggle with basic digital infrastructure. An over-reliance on AI-driven legal tools could alienate rural, economically backward, or linguistically diverse litigants.
Data Privacy Violations: Processing vast amounts of highly sensitive litigation data through external AI platforms presents severe data security hazards and risks non-compliance with the Digital Personal Data Protection (DPDP) Act, 2023.
The Burden of Verification: While the draft mandates a strict “Disclosure Regime” (requiring lawyers to declare AI use), enforcing this puts an extra administrative burden on courts to verify the factual authenticity of files, potentially adding delays.

Way Forward

Enforcing the “Human-in-the-Loop” Mandate: Ensure that AI remains strictly assistive. Core judicial responsibilities—such as evaluating witness credibility, determining bail, and sentencing—must remain the exclusive domain of human judges who possess emotional intelligence, equity, and contextual understanding.
Robust Institutional Oversight: Expedite the setup of the proposed permanent Apex Body at the Supreme Court level along with High Court AI committees. This body must establish strict evaluation standards, an “AI Incident Database,” and clear procurement policies for technology vendors.
Mandatory Audits and Sandboxing: Before any tool is deployed across the nationwide network of courts, it must undergo Controlled Environment Testing (sandboxing). Regular, independent algorithmic audits must be conducted to check for underlying biases, technical glitches, and security vulnerabilities.
Bridging the Digital Divide: Digital infrastructure in sub-district and rural courts must be upgraded systematically alongside the introduction of AI tools. AI applications should be focused on local language translation (e.g., maximizing the potential of tools like SUVAS) to actively democratize access to justice.
Capacity Building and Training: Legal professionals, court staff, and judicial officers need comprehensive training on the limitations of AI. Lawyers must be made strictly accountable for verifying AI-generated research, reinforcing that technology is an enabler, not a shield for professional negligence.

Preliminary Examination Question

Q. Consider the following statements regarding the draft ‘Regulations for Use of Artificial Intelligence (AI) in Courts, 2026’:

The regulations strictly prohibit the use of AI for administrative functions like translation of judgments and case scheduling.
The draft regulations mandate that data processing by AI systems in courts must comply with the Digital Personal Data Protection Act, 2023.
The proposed apex body for AI supervision in the judiciary will be chaired by the Union Minister of Electronics and Information Technology.

Which of the statements given above is/are correct?

A) 1 and 2 only

B) 2 only

C) 2 and 3 only

D) 1, 2 and 3

Answer: B) 2 only

Explanation: Statement 1 is incorrect because the draft explicitly permits AI for administrative functions like translation and case management. Statement 3 is incorrect because the proposed apex body will be chaired by a Supreme Court judge nominated by the Chief Justice of India, not the Union Minister. Statement 2 is correct as per the draft provisions.

Mains Examination Question

Q. “Technology can assist justice, but it cannot replace the judge.” In light of the draft ‘Regulations for Use of Artificial Intelligence (AI) in Courts, 2026’, critically evaluate the opportunities and challenges of integrating AI into the Indian judicial system. (15 Marks, 250 Words)

Key Pointers to Structure Your Answer:

Introduction: Mention the context of the draft Regulations, 2026, and the Supreme Court AI Committee’s stance on keeping AI strictly assistive.
Opportunities (The Assistive Role): Discuss how AI can help reduce the massive pendency of cases in India (e.g., automated case management, rapid translation of multilingual judgments, automated transcripts, and streamlined scheduling).
Challenges (The Core Threats): Highlight issues like AI hallucinations (fake precedents), lack of transparency in “opaque” algorithms, reinforcement of social biases, risk to personal liberty via biased “risk scoring” for bails, and alignment with data privacy laws (DPDP Act, 2023).
Analysis of the Draft: Appreciate the ‘Human-in-the-loop’ philosophy and the restriction on core judicial outcomes (sentencing, profiling).
Conclusion: Conclude with a balanced view stating that AI should be used as a tool to enhance efficiency, but the final ethical, moral, and legal calculus must always rest with human intellect.