Introduction
India’s judiciary is welcoming artificial intelligence, but only as an assistant, not as a judge. The Supreme Court’s draft regulations on AI in courts make clear that core adjudicatory powers must remain with human judicial officers, even as technology is used to improve efficiency, access, and case management.[1]
The issue is not whether AI can be useful. It already can. The real question is whether a machine can ever be trusted to exercise judicial discretion, weigh credibility, and deliver justice consistently with constitutional guarantees. In India, the answer is still no.
The Supreme Court’s approach
In June 2026, the Supreme Court of India placed draft regulations for the use of AI in courts in the public domain for consultation. The draft reflects a human-first model: AI may assist with tasks such as transcription, translation, research, scheduling, and document review, but it cannot decide cases, determine bail, assess witness credibility, or pronounce verdicts.[2]
That distinction matters. The draft also requires human oversight, impact assessments, and disclosure of AI use where relevant.[3] In other words, the Court is not rejecting technology; it is regulating it so that it serves judicial work without taking over judicial power. This is the central idea behind the proposed framework.
Constitutional limits
A separate law of logic governs the courtroom: justice must not only be done but must be seen to be done. If AI were allowed to replace judges, it would raise serious concerns under Articles 14 and 21 of the Constitution of India, which guarantee equality before the law and the right to life and personal liberty. A decision made by a black-box system that cannot clearly explain its reasoning would sit uneasily with both principles.
The Constitution also demands accountability. A judge can be questioned on legal reasoning, recusal, bias, or error. An algorithm cannot be cross-examined in the same way. That is why the Supreme Court’s emphasis on human primacy is constitutionally sound. It protects due process, transparency, and the right of litigants to understand the basis of decisions affecting their liberty, property, and dignity.
Case law support
Indian case law already points in this direction. In Selvi v State of Karnataka,[4] the Supreme Court held that involuntary techniques such as narco-analysis, polygraph tests, and brain mapping violated Article 20(3) and Article 21 because they compromised autonomy and fair procedure. the case is important here because it shows the Court’s reluctance to permit technological methods that weaken constitutional protections in the criminal process.
Similarly, in Justice KS Puttaswamy (Retd) v Union of India,[5] the Court recognised privacy as a fundamental right. That judgment is highly relevant to AI in the justice system because most AI tools rely on large volumes of personal and sensitive data. If such data is processed without proper safeguards, consent, and purpose limitation, privacy and dignity may be compromised.
The Court’s broader jurisprudence also recognises that functions requiring judgment and discretion cannot simply be handed over to non-human actors. In State of Punjab v Devans Modern Breweries Ltd,[6] the Court stressed limits on delegation where important legal authority is concerned. That principle supports the idea that judicial decision-making, which is at the heart of state power, cannot be outsourced to AI.
Benefits without replacement
It would be a mistake, however, to treat AI as a threat only. Properly used, it can make the justice system faster and more accessible. AI can help courts manage large case backlogs, transcribe proceedings, translate judgments into regional languages, and identify procedural defects early. Those uses do not replace judicial reasoning; they support it.
The Supreme Court’s draft regulations appear to accept this reality. They allow AI in administrative and assistive functions, while drawing a firm line at decision-making.[7] That is a sensible balance. Courts are under pressure to reduce delay, but speed cannot come at the cost of fairness. A system that is quicker but less just would fail its constitutional purpose.
Risks of machine justice
The biggest problem with replacing human decision-making is bias. AI systems learn from data, and data often reflects historical discrimination, error, or imbalance. If such data is used in judicial contexts, the result may be automated unfairness disguised as objectivity. A machine may look neutral, but its outputs can still reproduce existing prejudice.
There is also the problem of explainability. Many AI models operate as black boxes, meaning their internal reasoning is difficult to trace. In a court of law, that is unacceptable where liberty or liability is at stake. A litigant must be able to know why a decision was made, not merely that a system produced it. The draft regulations wisely reject undisclosed or unexplainable AI systems in adjudication.[8]
Another concern is overreliance. If judges begin to trust AI recommendations too much, human independence may slowly weaken. That danger is subtle but real. Judicial officers must remain critically engaged with the evidence and arguments before them, rather than allowing algorithmic suggestions to shape outcomes by default.
Statutory and policy context
India is already building a governance framework around AI. The Government’s India AI Governance Guidelines 2025 adopt a principle-based model focused on safety, accountability, and responsible use.[9] That framework supports the Supreme Court’s cautious approach and suggests that India is moving toward regulation rather than prohibition.
The Digital Personal Data Protection Act, 2023 also matters here because judicial AI tools may process sensitive personal information. If the justice system uses AI without strong data safeguards, compliance with privacy law becomes a major concern. The Court’s consultation process therefore sits within a wider regulatory environment, not in isolation.
Conclusion
AI should not replace human decision-making in the justice system. It can improve efficiency, reduce backlog, and support administrative work, but adjudication must remain human because fairness, accountability, and constitutional legitimacy depend on it. The Supreme Court’s draft regulations get that balance right by allowing AI as a tool while preserving the judge as the final decision-maker.[10]
For India, the best path forward is not machine justice, but human justice strengthened by carefully controlled technology. That approach preserves both innovation and the constitutional values on which the legal system rests.
Reference
[1] Supreme Court of India, Draft Regulations for Use of Artificial Intelligence (AI) in Courts, 2026.
[2] Supreme Court of India, draft regulations on AI in courts, public consultation notice, June 2026.
[3] Ibid.
[4] [2010] 7 SCC 263.
[5] [2017] 10 SCC 1.
[6] [2004] 11 SCC 26.
[7] Supreme Court draft regulations on AI in courts, 2026.
[8] Ibid.
[9] Government of India, India AI Governance Guidelines 2025.
[10] Supreme Court of India, draft regulations on AI in courts, 2026.