Editorial Summary: The Hindu argues that the rapid integration of Artificial Intelligence tools – the newly launched Su-Sahayak chatbot, SUPACE for case management and SUVAS for vernacular translation – into the Indian judiciary promises substantial relief from a 5.1-crore case backlog, but raises unresolved questions of data protection, algorithmic bias and constitutional accountability. The editorial cautions that without statutory safeguards under the Digital Personal Data Protection Act, 2023 and judicially developed transparency norms, AI in courts could erode the Article 21 guarantee of due process rather than enhance it.

A Backlog That Invited the Machine

As of early 2026, more than 5.1 crore cases are pending across the three tiers of Indian courts – about 80,000 in the Supreme Court, 62 lakh in High Courts and over 4.5 crore in subordinate courts, according to the National Judicial Data Grid (NJDG). The judge-to-population ratio remains close to 21 per million, far below the 50 per million benchmark recommended by the Law Commission of India in its 120th Report (1987). Against this overload, the judiciary has turned to AI as an administrative force multiplier rather than a decisional substitute.

The Supreme Court Portal for Assistance in Courts Efficiency (SUPACE), launched in April 2021 by then CJI S. A. Bobde, was the first significant experiment – an AI-driven research assistant designed to help judges read briefs, extract issues and locate precedent. The Supreme Court Vidhik Anuvaad Software (SUVAS) translates judgments into nine scheduled languages, addressing the long-standing barrier that 99 per cent of Supreme Court rulings are written only in English. Su-Sahayak, the AI chatbot launched in 2026 by CJI Surya Kant, is the latest addition – a citizen-facing tool that answers procedural queries and case-status questions in natural language.

The eCourts Foundation

These tools sit on top of the eCourts Mission Mode Project, now in its third phase. Phase III, approved by the Union Cabinet in September 2023 with an outlay of Rs 7,210 crore, runs from 2023 to 2027 and includes virtual courts, paperless filing, digital case management and machine-readable case data – the substrate that AI tools require to function. The NJDG, India Code and the eFiling portal have already digitised the front-end of justice delivery.

The promise is real. AI-assisted research can shorten judgment-drafting cycles, SUVAS expands access for non-English litigants, and Su-Sahayak removes the touts who long monopolised information about court procedure. But the architecture is being built faster than the safeguards.

Bias, Black Boxes and the COMPAS Warning

Algorithmic bias is not a hypothetical concern. The COMPAS recidivism prediction tool used in several US states was shown by a 2016 ProPublica investigation to systematically over-predict recidivism for Black defendants. The State v. Loomis (Wisconsin, 2016) decision – where COMPAS scores influenced sentencing despite the algorithm’s opacity – has become the cautionary template internationally.

India’s risks are not identical but are structurally similar. Training data drawn from past judgments reflects historical patterns of arrest, prosecution and sentencing, which themselves reflect caste, religion and class disparities. An AI tool trained on this corpus will, without conscious correction, encode those disparities into its outputs. The judiciary has so far not published the training datasets, model architectures or evaluation protocols for SUPACE or Su-Sahayak.

The DPDP Act and the Data Question

The Digital Personal Data Protection (DPDP) Act, 2023 came into force from 2025 with phased rule notifications. It governs personal data processing by both state and private fiduciaries, but Section 17(2)(a) carves out an exemption for the state in the interests of sovereignty, public order and judicial functions. AI tools running on litigant data, witness statements and case histories occupy precisely this grey zone – ostensibly a judicial function, but operationally a data-processing activity vulnerable to leakage, scraping and inference attacks. Courts have not yet clarified whether the Data Protection Board has jurisdiction over judicial AI systems.

UPSC Mains Analysis

GS Paper 2 – Governance and judiciary / GS Paper 3 – Awareness in IT, data protection

Key arguments:

  • SUPACE, SUVAS and Su-Sahayak together represent a credible workload-relief strategy for a 5.1-crore-case backlog and a structurally under-staffed judiciary.
  • The eCourts Phase III outlay of Rs 7,210 crore provides the digital infrastructure that AI in courts requires.
  • Algorithmic bias is a documented international risk (COMPAS, Loomis) and Indian training data carries structural inequities.
  • The DPDP Act, 2023 contains a state-function exemption that leaves judicial AI in a regulatory grey zone.

Counterarguments:

  • AI tools today assist, not decide – the constitutional locus of judgment remains with the bench, so concerns may be premature.
  • Vernacular translation through SUVAS democratises access for non-English litigants, a clear Article 14 gain.
  • Without AI assistance, the alternative is continued backlog and denial of timely justice – itself an Article 21 violation.

Keywords: SUPACE, SUVAS, Su-Sahayak, eCourts Phase III, NJDG, DPDP Act 2023 Section 17, Article 21 due process, algorithmic bias, COMPAS, State v. Loomis, Law Commission 120th Report, CJI Surya Kant.

Editorial Insight

The Hindu’s view is that AI in Indian courts is neither saviour nor saboteur but a powerful instrument that demands constitutional discipline. The judiciary cannot demand transparency from the executive while running its own AI systems as black boxes. The next step is not more tools but a published framework – training data audits, bias-evaluation protocols, an independent judicial-AI ethics committee, and a clear answer to who is accountable when the algorithm errs. Otherwise the price of clearing the docket will be paid in due process.