I. Introduction
Conflict is not an aberration in human affairs. It is a constant. Quarrels between neighbours, traders, communities and nations fill the pages of our history, and every civilisation has built machinery to resolve Beneath every such quarrel lies the same demand, that each person should receive what is due to them. The Roman jurists gave that demand its enduring form, defining justice as “the constant and perpetual wish to render every one his due”.[1] That will has not changed. The machinery has, and as societies and commerce developed, it has had to be rebuilt time and again.
For centuries, the court as we know it has been the central piece of that machinery, and it has been understood above all as a place.[2]The litigant travelled to the building, stood before the judge and received the decision in the room where it was made. Professor Richard Susskind has drawn an important distinction here – he asks whether a court is a service or a [3] I second his view that our understanding of justice has been tied to a place, when it should be understood as a service. If justice is the constant and the court only the machinery, the building is a means and not the end. A courtroom is one way of delivering that service. It is not the only way, and a person who cannot reach the building should not, for that reason, be denied what the building exists to provide.
The need for that shift is pressing, because a justice system tied to a place leaves vast numbers of people outside it. The most comprehensive global estimate, published in 2019, found that 5.1 billion people, about two-thirds of the world’s population, faced at least one justice need they could not resolve.[4]The World Justice Project’s 2023 survey of 104 countries places 4 billion people in the justice gap on account of civil and administrative problems alone.[5] The OECD reported in 2025 that access to justice has declined in most countries since 2016, and that only 54% of citizens in its member States trust their justice system.[6]
In India, the gap takes several forms. The most visible is pendency. The Government informed the Rajya Sabha in July 2026 that 5.64 crore cases were pending across the country, with nearly nine in ten before the district and subordinate courts.[7]Behind the pendency lies the capacity of the system. The India Justice Report 2025 found roughly 15 judges for every million people, against the Law Commission’s recommendation of 50, with a third of High Court posts lying vacant.[8] For many people, distance and cost keep them from the courts long before their cases are ever counted as pending. On figures of this order, the maxim that justice delayed is justice denied ceases to be a saying and becomes a description.
Technology has made the shift from place to service possible, and it has also changed the disputes themselves. Commerce is conducted on platforms, contracts are concluded by an exchange of messages, and payments move in seconds, often between parties in different The disputes that arise from all of this arrive in numbers, in forms and across borders that the traditional courtroom was never designed to handle. Information and communication technology, and now artificial intelligence (“AI”), are altering how disputes arise, how they are resolved and how those who resolve them do their work. This is the changing landscape my address concerns.
Technology is also the most realistic means of closing the gap I have described, and the Indian legal system recognised that early. The e-Courts project, conceived in 2005 and launched in 2007, was the institutional response to precisely this problem.[9]The pandemic then turned a gradual transition into an urgent one. Courts, tribunals and mediators adapted within weeks, and what they learnt in those months has stayed with them.
Two decades into that transition, and with innovation now arriving faster than ever, four questions arise, and they form the subject of this address:
(i) What has technology already changed in the way disputes are resolved?
(ii) Does it merely make the old process faster, or does it alter the process itself?
(ii) What does it put at risk?
(ii) How should the legal system proceed, so that the gains are kept and the risks contained?
My address takes up these questions in turn. Part I traces how technology has automated existing processes in courts and in alternative dispute resolution (“ADR”). Part II turns to online dispute resolution (“ODR”), where technology begins to redesign the process Part III examines the challenges that these changes bring. Part IV sets out the principles that should guide the way forward. Finally, I will conclude with a few observations on where these changes may take the legal system.
II. Part I: Technology in Courts and ADR: Automating the Existing Process
So what has technology already changed? The answer begins with the law, because courts and Parliament first had to accept that justice could be administered through technology. It then turns to practice, where the change has taken different routes in courts and in In the courts, it has been driven by a national programme. In arbitration and mediation, it has been driven by the parties and the institutions that serve them.
(i) Recognition in Law
Justice P.N. Bhagwati reminded us that law is not a mausoleum. It is “essentially a social process, the end product of which is justice”, and it must keep growing with the society it serves.[10]The law relating to dispute resolution has done exactly that, first through the courts; in Grid Corporation of Orissa Ltd v AES Corporation (2002), the Supreme Court accepted that arbitrators could consult one another through electronic media.[11] A year later, in State of Maharashtra v Praful B Desai (2003), it held that evidence could be recorded by video conference.[12] By 2018, it had gone further. In Meters and Instruments Pvt Ltd v Kanchan Mehta (2018), it observed that some categories of cases could be dealt with online without the physical presence of the parties, and in Swapnil Tripathi v Supreme Court of India (2018), it held that proceedings of public importance should be streamed live.[13]
Parliament then gave this judicial recognition a statutory footing. The Arbitration and Conciliation Act, 1996 recognises an arbitration agreement contained in an exchange of electronic communications, and in Belvedere Resources DMCC v OCL Iron and Steel Ltd (2025), the Delhi High Court applied that provision to an agreement concluded over WhatsApp and The Mediation Act, 2023 expressly provides for online mediation.[14]In other words, the law did not merely tolerate technology. It progressively recognised technology as a legitimate means of administering justice.
(ii) Technology in the Courts
Recognition made technology-enabled justice possible. Delivering it required institutional capacity, and in the courts, the e-Courts project built that capacity in three phases. The first, completed in 2015, computerised more than 14,000 courts. The second, from 2015 to 2023, introduced the National Judicial Data Grid (“NJDG”), electronic filing and e-Sewa Kendras, and expanded video conferencing more than fivefold. The third, which runs from 2023 to 2027 with an outlay of INR 7,210 crore, aims to make courts paperless and [15]The results are best understood by what they have changed for the litigant.
(i) Records and filing: Electronic filing has enabled the submission of more than 1.25 crore cases, and the NJDG makes the status of cases
(ii) Participation: Courts have conducted more than 18 crore hearings by video conference, witnesses and accused persons can now testify remotely, and Constitution Bench hearings have been streamed live since September 2022.
(iii) High-volume matters: Virtual courts have received 11.33 crore traffic challans without a single litigant entering a courtroom.
(iv) Integration: The Interoperable Criminal Justice System allows the police, courts, prisons, forensic laboratories and prosecution to exchange records electronically, and e-Sakshya captures digital evidence at the [16]Tools such as Nyaya Setu, the multilingual AI assistant under the DISHA programme, can give a person plain information about rights and remedies in their own language before a grievance hardens.
AI has entered the same courts, so far in an assisting role, and its uses have grown from administrative to intellectual. The Supreme Court began with translation. With the help of SUVAS, a machine-learning tool, it has now published more than 83,000 judgments in 18 Indian languages.[17]Since February 2023, TERES has transcribed arguments before the Constitution Benches live and published the transcripts.[18] The same tool has since been approved for real-time transcription in Singapore courts.[19] Other tools convert dictation into text, translate orders and scrutinise electronically filed petitions for [20] On the intellectual side, SUPACE was launched in April 2021 to collect relevant facts and law for judges, and a research assistant, LegRAA, is now being tested.[21] At the launch of SUPACE, Chief Justice Bobde made clear that the system was designed to assist judges and not to decide. That limit has held so far, and it will matter when I turn to the challenges.
(iii) Technology in Arbitration and Mediation
Since I demitted office as a judge of the Supreme Court in 2019, much of my work has been in arbitration and mediation, and much of the change I have witnessed there has come from Arbitration and mediation have covered similar ground to the courts, but by a different route. No national programme directs them, because parties and institutions are free to adopt whatever serves them, and change has therefore come from the users. TERES itself was first used by arbitration practitioners, and the Supreme Court took it up after its value had been shown there.[22]That is a small but telling instance of ADR testing a tool before the courts adopt it.
Virtual hearings followed a similar When the 2021 Queen Mary and White & Case survey asked users how they wished to proceed after the pandemic, respondents preferred a mix of in-person and virtual formats for almost all types of interaction, while recording concern about assessing the credibility of witnesses on screen.[23]My own practice bears this out. Procedural hearings are now routinely held online, and the parties assemble in person only for evidence and final arguments. Online mediation, now recognised by statute, offers the same convenience. The time and expense saved to the parties is considerable, and it has come without any loss in the quality of the hearing that matters.
AI has followed in The 2025 edition of the same survey found that the principal drivers of AI use in arbitration were the saving of time, at 54%, and the reduction of cost, at 44%.[24]Anyone who has sat in a construction dispute knows that pleadings and annexures run into thousands of pages, and a tool that indexes and cross-references them clears the ground on which the tribunal decides. Arbitral institutions have issued guidance on such use, and that guidance requires arbitrators to retain their independent judgment.[25]
(iv) From Automation to Innovation
All of this is significant, but it has one thing in Each tool helps an existing process to run faster. A suit is still filed, heard and decided, and an arbitration still moves from pleadings to award. That is automation, which means doing the same things more efficiently. Innovation is different. It asks whether the process itself can be redesigned, so that disputes are resolved, or even avoided, in ways the traditional process never allowed.[26]The two are not always easy to separate. Kerala’s 24×7 ON Court in Kollam, which began functioning in November 2024 for cheque dishonour cases, is still a court, yet it accepts filings at any hour and handles each case digitally from filing to disposal.[27] It stands somewhere between the two. ODR shows most clearly what innovation looks like, and that is why it deserves separate treatment.
III. Part II: Beyond Automation: Online Dispute Resolution
What, then, does it mean for technology to change the process itself, rather than speed it up? ODR supplies the clearest answer. Four aspects deserve attention: what ODR is, how India has adopted it, how it can prevent disputes, and what it means for the shape of the legal system.
(i) What ODR Is
ODR is often mistaken for mediation conducted over video. It is broader than that. It uses technology to assist negotiation, mediation and arbitration, separately or in combination, through a process designed for the kind of dispute it is meant to resolve. The difference is clearest at the point where ODR began. The traditional model asks a party with a grievance to leave the transaction and approach an external forum after the dispute has arisen. ODR reverses that order by building the remedy into the transaction itself. eBay, one of the earliest online marketplaces, illustrates why this became necessary. It could not send millions of small disagreements between buyers and sellers to court when a suit would cost more than the goods. It therefore built a process into its own platform, beginning with guided negotiation and moving to a neutral only if negotiation failed. By 2010, it was resolving about 60 million disputes a year in that way.[28]The tools have since In blind bidding, for instance, each party enters a confidential figure, and the system declares a settlement when the two figures come within an agreed range, so that neither side has to make the first concession openly.
(ii) India’s Adoption of ODR
India has taken this idea into public policy. In June 2020, at the height of the pandemic, NITI Aayog constituted a committee to prepare an action plan for mainstreaming ODR, and I had the privilege of chairing it. We reported in November 2021 and recommended measures at three levels.[29]
(i) At the structural level, we recommended better access to digital infrastructure, greater digital literacy, and training for professionals as neutrals.
(ii) At the behavioural level, the adoption of ODR for disputes involving Government departments and ministries.
(iii) At the regulatory level, a light-touch approach founded on design and ethical principles, so that a young sector could regulate itself and was not constrained prematurely.
ODR has since moved from a policy recommendation to an institutionalised part of dispute resolution in particular sectors.
(i) Payments: In August 2020, the Reserve Bank of India required payment system operators to provide ODR for failed transactions.[30]
(ii) Securities: In 2023 the Securities and Exchange Board of India introduced online conciliation and online arbitration across the securities market, conducted through empanelled ODR institutions.[31]
(iii) Small enterprises: The MSME Samadhaan portal had received 2,56,892 applications by December 2025, involving claims of more than INR 55,000 crore, and a dedicated MSME ODR portal was launched in June [32]
In each of these sectors, ODR has become part of the design of the transaction itself, so that a grievance meets a remedy at the point where it arises.
Two recommendations of the Committee deserve to be reiterated, because their importance has grown since. The first is that the Government, as the country’s largest litigant, should lead by resolving its own disputes online. A State that settles its contractual and service disputes in this way relieves the courts of a large share of their work and shows every other litigant that the process can be The second is that technology should be built into pre-litigation processes. Parliament has since made pre-institution mediation mandatory for commercial suits that do not contemplate urgent relief, and the Supreme Court held in Patil Automation Pvt Ltd v Rakheja Engineers Pvt Ltd (2022) that the requirement must be complied with.[33]That mandate places a large volume of commercial disputes before a mediator before they can reach a court. It can work at that scale only if a mediation can be convened within days, without the parties having to travel, and at a cost proportionate to the claim. An online process provides exactly that. Pre-litigation mediation and ODR are therefore not separate reforms. The first creates the obligation, and the second supplies the means of meeting it. Neither will achieve much, however, without a change in the mindset of parties and their advisers. So long as pre-litigation mediation is treated as a formality to be completed before the suit, no platform, however efficient, will produce settlements.
(iii) ODR as a Means of Dispute Prevention
The sectoral schemes, and the role of ODR in pre-litigation mediation, concern disputes that have already That is only part of its potential. Its deeper significance is that it can reach disputes before they arise. Professor Susskind describes access to justice in four layers. These are (i) the promotion of legal health, (ii) the avoidance of disputes, (iii) the containment of disputes and (iv) their authoritative resolution.[34] Courts work at the fourth layer, and ADR largely at the third. ODR today works mainly at the third, through online negotiation and mediation, and at the fourth, through online arbitration. Its larger promise lies at the first two layers, and that promise is still being realised. ODR platforms can give a person plain information about rights and remedies before a grievance hardens. In time, the data from thousands of similar complaints can show a business or a Government department where its disputes arise, so that the cause is removed rather than the dispute merely settled. In Professor Susskind’s image, it is better to have a fence at the top of the cliff than an ambulance at the bottom.[35]
(iv) What These Changes Mean for the Legal System
Taken together, the changes described in Parts I and II point to a legal system whose shape is altering. ODR does not displace the courts or traditional ADR. It complements them. A consumer complaint may be settled on a platform within days, a commercial dispute may go to online arbitration before an expert in another country, and the courts remain free for matters that require an authoritative and public ruling. The second example carries a further consequence. When parties, counsel and tribunal no longer need to share a city, they no longer need to share a country either, and a cross-border dispute can be resolved online as readily as a domestic one. That is why international standards, such as the UNCITRAL Technical Notes on Online Dispute Resolution, matter as much as domestic rules.[36]
The roles of those who work in the system are changing with it. A judge’s work has always involved examining the record, identifying the issues, researching the law and deciding. Technology can increasingly assist with the first three, though always under the judge’s supervision. The judge’s role will therefore centre more and more on what only a judge can do, namely weighing competing arguments, exercising judgment on contested questions and ensuring that the process is fair. The responsibility for the decision remains personal. The advocate’s role is shifting in the same direction. Less time will go into routine review and research, and more into verification, strategy and advocacy. A legal system built in this way succeeds not only by deciding disputes well but by reducing the number that need to be decided.
IV. Part III: The Challenges
The greater the reach of technology, the greater the responsibility to examine what it puts at risk. What, then, are those risks? They fall under three heads. The first is the legitimacy and integrity of justice, which depends on how proceedings are seen, how reliable the tools are and who decides. The second is the security and confidentiality of proceedings. The third is the accessibility and equality of justice. I take them in that order.
(i) The Legitimacy and Integrity of Justice
(a) Public Perception and Trial by Media
Lord Hewart’s dictum that justice should “manifestly and undoubtedly be seen to be done” is perhaps the best-known sentence in the law of procedure.[37]Live streaming was introduced to give effect to it. Technology has, however, changed how justice is seen. The public rarely watches a hearing from beginning to end. It receives a clip of a few seconds, selected and captioned by someone else and often monetised. Judicial proceedings have thus become material for public consumption outside their original context.
Trial by media is not new, but technology has changed its reach. Chief Justice Ramana observed in 2022 that the media had begun to run kangaroo courts on questions that even experienced judges find difficult.[38]Social media goes further than broadcast, because it combines enormous capacity to amplify with no means of separating the true from the fabricated. Public discussion of a case once followed the It now precedes it, and often concludes before the court has taken up the matter. Much of it is not even spontaneous. One study of more than 750,000 messages about a former Chief Justice of India, posted over 110 days in 2023, traced coordinated campaigns behind the surges of hostile comment.[39]
Why does this matter for the future of justice? There are three
(i) It changes how the public understands A question a judge asks to test a submission is circulated as the court’s view, and in a polarised debate the outcome of a case is read politically before it is read legally.
(ii) It changes the judge’s position. Criticism of a judgment is legitimate and healthy, but attacks on the person who wrote it now extend to threats against judges and their [40]A judge cannot answer, and a judiciary that begins to decide with one eye on the reception of a judgment is no longer deciding according to law alone.
(iii) It threatens the record itself, because tools now exist that can alter the words a speaker appears to say on video.
The answer is not to close the doors that technology has opened. An institution that opens itself must also regulate what is done with that opening, and the Supreme Court has begun to do so. In Harshita Grover v Union of India (2026), by an interim order of 24 July 2026, it directed that recordings of judicial proceedings shall not be extracted, edited, circulated or monetised on digital platforms without the prior permission of the Secretary General of the Supreme Court or the Registrar General of the High Court concerned, while leaving recognised news organisations free to report.[41] The matter is pending, and the Court has asked the Union Government to address enforcement in the policy it is framing.[42] Regulation alone will not suffice. The Bar, which was present in the courtroom, must be willing to explain what was actually said and decided when the judge cannot.
(b) Reliability of Technology and Digital Evidence
The second risk concerns what machines tell us, and it was visible well before AI arrived. For about fifteen years, the Post Office in the United Kingdom prosecuted more than 700 sub-postmasters based on shortfalls shown by its Horizon accounting software. The law presumed that a computer was working correctly unless the contrary was shown, and the accused had no means of showing it. The Court of Appeal quashed thirty-nine of those convictions in 2021, and Parliament quashed most of the rest by statute in [43]That was automation, and a fairly simple system at that. The lesson is that trust in a machine must be earned by testing and not presumed.
Generative AI magnifies the same risk, because its errors look exactly like correct answers, and it now threatens the integrity of legal reasoning In Pooja Ramesh Singh v Jammu and Kashmir Bank Ltd (2026), the NCLT decided an application under Section 7 of the Insolvency and Bankruptcy Code on the strength of precedents that did not exist. Counsel had not cited them. They entered the order through the tribunal’s own research, so the failure lay with the adjudicator and not the Bar, and the appellate tribunal affirmed the order without noticing. Setting both orders aside, the Supreme Court held that a decision resting even in part on fabricated authority cannot stand, and that human control must remain the governing principle of adjudication.[44]
A related risk concerns the integrity of the evidentiary record. A recording, an image or a document can today be fabricated or altered with little skill, and its authenticity can no longer be assumed. Advocates will have to ask where the material placed in their hands has come from. Courts and tribunals, for their part, would do well to require disclosure before the hearing of whether any recording or image has been enhanced or altered, so disputes about authenticity are resolved in advance, not in the middle of the
(c) Adjudication by AI
The third risk is the one the legal profession is now openly debating, namely whether AI should decide disputes itself. The question is no longer theoretical. In November 2025, the AAA-ICDR introduced a system that drafts awards in documents-only construction disputes of low value, for review and signature by a human arbitrator, where both parties agree.[45] Those who favour going further argue that a litigant wants an outcome, not a particular method of reaching it, and that human judges are not fully transparent about how they decide.[46] The question has already reached the courts. In LaPaglia v Valve Corporation (2025), a party asked a court in the United States to vacate an award on the ground that the arbitrator had outsourced his reasoning to AI. The court dismissed the petition on jurisdictional grounds without reaching the merits, but the challenge shows what has become arguable.[47] The real danger lies in transferring adjudicative authority, gradually and without deliberate choice, to a system whose reasoning cannot be examined and which cannot be held to account.
My own position is that we must build AI for adjudication, and not AI as adjudicator. Adjudication is the exercise of authority by a person who has heard the parties, applied their own mind and given reasons that can be tested on appeal, and who remains answerable for the The reasons a judge gives may not disclose every influence on the decision, but they are public, attributable to a named person, and can be corrected. A system whose reasoning cannot be examined offers the outcome without that discipline. Arbitration rests on the same principle, because the mandate of an arbitrator is personal. The limit stated at the launch of SUPACE, that the system assists and does not decide, is the right one, and it must hold as the tools grow more capable.
(ii) The Security and Confidentiality of Proceedings
The digitisation of proceedings creates new obligations. Courts, arbitral institutions, mediators and ODR platforms now hold pleadings, evidence, financial records and settlement discussions in digital The Digital Personal Data Protection Act, 2023 treats those who determine how personal data is processed as data fiduciaries and requires them to limit its use to the purpose for which it was collected, protect it with reasonable security safeguards, and report breaches. The legal position of courts and arbitral bodies is not identical. The Act exempts courts and tribunals in the performance of judicial functions, but the position of arbitral tribunals and institutions remains unsettled.[48] Confidentiality is also exposed to a more direct kind of disruption. Hearings before the Karnataka High Court in 2023 and the Delhi High Court in 2026 had to be suspended after intruders displayed obscene material on the video link.[49] Security must therefore be built into the design of digital proceedings, and not added after a breach. Confidentiality and order, which in a physical proceeding were secured by a closed door, must now be secured by encryption, authentication and audit.
(iii) The Accessibility and Equality of Justice
This address began with a justice gap caused by distance, cost and the limited capacity of courts. Technology reduces those barriers, but it can raise new ones in their place. The International Telecommunication Union estimates that 6 billion people used the internet in 2025, which leaves 2.2 billion who did not. In urban areas, 85% of people are online, against 58% in rural areas.[50]In the legal system, the divide concerns devices and connectivity, digital literacy, language, and disability.
The divide is ultimately a question of procedural A technologically efficient process is not necessarily a fair one. An online hearing may be quicker, yet it may deprive a party of a real opportunity to present its case if that party lacks the means or the help to take part. AI adds a further dimension, because the best tools carry a cost. If one party can afford them and the other cannot, the equality of arms on which adversarial procedure depends is disturbed. The novelist William Gibson is credited with the remark that the future is already here, and is only not very evenly distributed.[51]That is an accurate description of digital justice today. The tools exist, but those who most need them are often the least able to use them.
V. Part IV: The Way Forward
How, then, should the legal system proceed? Four principles follow, and each answers one of the challenges I have described.
(i) Human primacy, which answers the risk of AI Technology may assist every function in the legal system, but the decision must remain that of a judge, an arbitrator or, in a settlement, the parties themselves. The draft Regulations for Use of Artificial Intelligence in Courts, 2026, published by the AI Committee of the Supreme Court in June 2026, rest on exactly this principle. They would permit AI to assist the work of courts while prohibiting its use to reach decisions.[52]
(ii) Accountability addresses the risk of unverified output. Responsibility cannot be transferred to a The judge answers for the judgment and the arbitrator for the award, and the advocate answers for every authority and every document placed before the forum, whatever tool helped to produce it. The output of any tool should therefore be treated as a senior would treat the draft of a capable but inexperienced junior. It may be used, but only after it has been read, checked against the record and adopted as one’s own.
(iii) Inclusive access, which answers the digital Digital justice must be designed for the least equipped user and not the best equipped. That requires services in the languages people speak, assisted services such as the e-Sewa Kendras, of which more than 1,800 are now fully functional in court complexes, and a physical alternative that remains open as of right.[53]
(iv) Institutional capacity, which addresses the risks to security and the gaps in readiness. It has two parts. The first is preparedness. A UNESCO survey of judicial operators in 96 countries found that 44% were already using AI tools in their work, while only 9% had received any institutional guidance or training.[54]Judges and court staff therefore need training before the tools arrive, not after, and the Bar and law schools must treat technological competence, including verifying its output, as part of professional competence. The second is governance. Arbitral institutions and ODR platforms must maintain transparent, secure processes, and their rules should specify when a party’s or tribunal’s use of AI must be disclosed. Policymakers should prefer principles and standards that can keep pace with the technology over detailed rules that it will soon outrun, which is the light-touch approach the NITI Aayog Committee recommended for ODR.
One standard governs all four principles, and it is the standard by which the future of justice should be judged. Efficiency must not be mistaken for justice. A hearing may be quicker without being more A platform may dispose of more disputes without resolving them fairly. A tool may produce an answer in seconds that is neither reliable nor sound in law. Speed is a means, and the legal system must keep asking whether it serves the end.
VI. Conclusion
This address began with the Roman jurists’ definition of Nothing in the changing landscape alters the will it describes. What has changed, and will keep changing, is the machinery. Hearings are held across continents, disputes are settled on platforms without a hearing at all, and machines now assist with work once done only by lawyers and judges.
Where may these changes take us? If the principles I have described hold, the legal system of the coming decade may look quite different from the one we Routine disputes may increasingly be resolved, and many avoided, outside formal hearings. Courts may be reserved more and more for the matters that require authoritative adjudication. Arbitration may become faster and less costly, with tribunals assisted at every stage by tools that do not decide for them. The larger aspiration, and it remains an aspiration, is that a litigant in a district town should come to have the same access to a neutral as a corporation in a metropolis.
Technology can carry justice to far more people than the courtroom alone ever could, and it can help to prevent disputes as well as resolve them. It cannot supply the qualities that make justice worth having, which are fairness, independent judgment and accountability. Those remain the responsibility of judges, arbitrators, mediators and the Bar. The future of
justice will depend on whether we use the new machinery in the service of that old will, and not the other way around.
*Justice AK Sikri, International Judge: Singapore International Commercial Court and Former Judge Supreme Court of India
[1] Justinian, Institutes 1.1 pr, in Thomas Collett Sandars (tr), The Institutes of Justinian (5th edn, Longmans, Green and Co 1874).
[2] Richard Susskind, Online Courts and the Future of Justice (Oxford University Press 2019).
[3] ibid; Richard Susskind, ‘The Future of Courts’ (2020) 6(5) The Practice https://clp.law.harvard.edu/knowledge-hub/magazine/issues/remote-courts/the-future-of-courts/.
[4] World Justice Project, Measuring the Justice Gap: A People-Centered Assessment of Unmet Justice Needs Around the World (June 2019) https://worldjusticeproject.org/our-work/research-and-data/measuring-justice-gap.
[5] World Justice Project, Dissecting the Justice Gap in 104 Countries: WJP Justice Data Graphical Report I (2023) https://worldjusticeproject.org/wjp-access-to-justice-report-part-1.
[6] OECD, Making Justice Systems More Effective and People-Centred (OECD Publishing 2025) https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/11/making-justice-systems-more-effective-and-people-centred_d705c01f/e02fd90b-en.pdf.
[7] Ministry of Law and Justice, Written Reply in the Rajya Sabha (23 July 2026), reported in Press Trust of India, ‘More than 5.64 Crore Cases Pending across Indian Courts: RS’ Telangana Today (23 July 2026) https://telanganatoday.com/more-than-5-64-crore-cases-pending-across-indian-courts-rs.
[8] Tata Trusts and others, India Justice Report 2025 (April 2025), reported in Press Trust of India, ‘Only 15 Judges per Million Population in the Country: 2025 India Justice Report’ ThePrint (15 April 2025) https://theprint.in/india/only-15-judges-per-million-population-in-the-country-2025-india-justice-report.
[9] Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf.
[10] Motilal Padampat Sugar Mills Co Ltd v State of Uttar Pradesh (1979) 2 SCC 409.
[11] Grid Corporation of Orissa Ltd v AES Corporation (2002) 7 SCC 736.
[12] State of Maharashtra v Praful B Desai (2003) 4 SCC 601.
[13] Meters and Instruments Pvt Ltd v Kanchan Mehta (2018) 1 SCC 560; Swapnil Tripathi v Supreme Court of India (2018) 10 SCC 639.
[14] Arbitration and Conciliation Act 1996, s 7(4)(b); Belvedere Resources DMCC v OCL Iron and Steel Ltd 2025:DHC:5128; Mediation Act 2023, s 30.
[15] Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf; Ministry of Law and Justice, Reply to Lok Sabha Unstarred Question No 856 (26 July 2024) https://eparlib.nic.in/bitstream/123456789/2977904/1/AU856_o5fVc1.pdf.
[16] Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf. The backgrounder also records that more than 753 crore pages of court records have been digitised, that 31 virtual courts are functioning, and that the Nyaya Shruti system allows accused persons, witnesses and experts to testify by video.
[17] ibid; Supreme Court of India, Indian Judiciary: Annual Report 2024-25 (2025) 219-225.
[18] ‘Breaking: Supreme Court Utilises AI to Make Live Transcriptions of Its Hearings’ LiveLaw (21 February 2023) https://www.livelaw.in/top-stories/supreme-court-ai-live-transcriptions-of-its-hearings-222108.
[19] S N Thyagarajan, ‘TERES Joins Singapore Courts’ Approved List for Real-Time Transcription Services’ Bar and Bench (21 September 2026) https://www.barandbench.com/news/law-policy/teres-joins-singapore-courts-approved-list-for-real-time-transcription-services.
[20] Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf. The tools are ASR-SHRUTI, PANINI and a defect-flagging tool developed with IIT Madras, respectively.
[21] ‘Won’t Let Artificial Intelligence Do Decision Making; Judges’ Autonomy and Discretion Will Be Retained: CJI Bobde’ LiveLaw (6 April 2021) https://www.livelaw.in/top-stories/supreme-court-artificial-intelligence-portal-supace-chief-justice-sa-bobde-172220; Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf.
[22] ‘Supreme Court Hearings to Be Transcribed Live’ Bar and Bench (21 February 2023) https://www.barandbench.com/news/supreme-court-hearings-to-be-transcribed-live.
[23] Queen Mary University of London and White & Case, 2021 International Arbitration Survey: Adapting Arbitration to a Changing World (2021) https://arbitration.qmul.ac.uk/research/2021-international-arbitration-survey/.
[24] Queen Mary University of London and White & Case, 2025 International Arbitration Survey: The Path Forward: Realities and Opportunities in Arbitration (2025).
[25] Chartered Institute of Arbitrators, Guideline on the Use of AI in Arbitration (March 2025) https://www.ciarb.org/media/m5dl3pha/ciarb-guideline-on-the-use-of-ai-in-arbitration-2025-final_march-2025.pdf; AAA-ICDR, Guidance on Arbitrators’ Use of AI Tools (March 2025).
[26] Richard Susskind, ‘The Future of Courts’ (2020) 6(5) The Practice https://clp.law.harvard.edu/knowledge-hub/magazine/issues/remote-courts/the-future-of-courts/; Richard Susskind, How to Think About AI: A Guide for the Perplexed (Oxford University Press 2025).
[27] ‘India’s First 24×7 Online Court Launched in Kollam, Kerala’ Bar and Bench (20 November 2024) https://www.barandbench.com/news/indias-first-24×7-online-court-launched-in-kollam-kerala.
[28] NITI Aayog Expert Committee on ODR, Designing the Future of Dispute Resolution: The ODR Policy Plan for India (November 2021) ch II https://www.niti.gov.in/sites/default/files/2023-03/Designing-The-Future-of-Dispute-Resolution-The-ODR-Policy-Plan-for-India.pdf.
[29] ibid.
[30] Reserve Bank of India, ‘Online Dispute Resolution (ODR) System for Digital Payments’ (Circular, 6 August 2020).
[31] Securities and Exchange Board of India, Circular No SEBI/HO/OIAE/OIAE_IAD-1/P/CIR/2023/131 (31 July 2023).
[32] Ministry of Micro, Small and Medium Enterprises, ‘MSME Samadhaan Portal Data’ (December 2025).
[33] Commercial Courts Act 2015, s 12A; Patil Automation Pvt Ltd v Rakheja Engineers Pvt Ltd (2022) 10 SCC 1.
[34] Richard Susskind, Online Courts and the Future of Justice (Oxford University Press 2019); NITI Aayog Expert Committee on ODR, Designing the Future of Dispute Resolution: The ODR Policy Plan for India (November 2021) 7-8.
[35] Richard Susskind, Online Courts and the Future of Justice (Oxford University Press 2019).
[36] UNCITRAL, Technical Notes on Online Dispute Resolution (United Nations 2017) https://uncitral.un.org/sites/uncitral.un.org/files/media-documents/uncitral/en/v1700382_english_technical_notes_on_odr.pdf.
[37] R v Sussex Justices, ex p McCarthy [1924] 1 KB 256 (KB) 259.
[38] K Rajagopal, ‘Media Is Running Kangaroo Courts: CJI Ramana’ The Hindu (23 July 2022) https://www.thehindu.com/news/national/media-is-running-kangaroo-courts-cji/article65673959.ece.
[39] Joyojeet Pal and others, ‘Hate Speech and Coordinated Campaigns against the Judiciary in India’ (University of Michigan School of Information 2023) https://joyojeet.people.si.umich.edu/?p=1165. The study found that the largest accounts joined each surge of interest a day or two after it began.
[40] ‘Retired Bombay HC Judge GS Patel, Family Face Threats and Attack over Dawoodi Bohra Succession Verdict’ LiveLaw (8 June 2026) https://www.livelaw.in/top-stories/retired-bombay-hc-judge-gs-patel-family-face-threats-and-attack-over-dawoodi-bohra-succession-verdict-537106.
[41] Harshita Grover v Union of India WP (C) No 751 of 2026 (SC, 24 July 2026).
[42] ‘Supreme Court Asks MeitY to Include Action against Unauthorised Court Hearing Clips in Proposed Policy’ LawChakra (October 2026) https://lawchakra.in/supreme-court/court-hearing-clips-meity/.
[43] Hamilton v Post Office Ltd [2021] EWCA Crim 577; Post Office (Horizon System) Offences Act 2024.
[44] Pooja Ramesh Singh v Jammu and Kashmir Bank Ltd 2026 INSC 668.
[45] AAA-ICDR, ‘Announcement on the Launch of its AI Arbitrator’ (3 November 2025).
[46] Richard Susskind, How to Think About AI: A Guide for the Perplexed (Oxford University Press 2025).
[47] LaPaglia v Valve Corporation No 3:25-cv-00833 (SD Cal, 9 December 2025).
[48] Digital Personal Data Protection Act 2023, ss 2(i), 8 and 17(1)(b); Mayank Grover and Pratibha Vyas, ‘India’s Data Protection Turn and the Future of Arbitral Confidentiality: Lessons from the EU and Singapore’ (Legal 500, 3 November 2026) https://www.legal500.com/intelligence/india/litigation-mediation-arbitration/indias-data-protection-turn-and-the-future-of-arbitral-confidentiality-lessons-from-the-eu-and-singapore.
[49] ‘Karnataka High Court Halts Online Hearings as Miscreants Air Pornographic Content during Proceedings’ Scroll (5 December 2023) https://scroll.in/latest/1060178/scroll_in; ‘Pornographic Content Disrupts Delhi High Court Virtual Hearing; Probe Ordered’ The Tribune (29 April 2026) https://www.tribuneindia.com/news/delhi/pornographic-content-disrupts-delhi-high-court-virtual-hearing-probe-ordered/.
[50] International Telecommunication Union, Measuring Digital Development: Facts and Figures 2025 (ITU 2025) https://www.itu.int/en/mediacentre/Pages/PR-2025-11-17-Facts-and-Figures.aspx.
[51] The remark is traced to William Gibson’s interview on NPR, Talk of the Nation (30 November 1999): ‘The Future Has Arrived: It’s Just Not Evenly Distributed Yet’ (Quote Investigator, 24 January 2012) https://quoteinvestigator.com/2012/01/24/future-has-arrived/.
[52] Supreme Court of India AI Committee, Draft Regulations for Use of Artificial Intelligence (AI) in Courts, 2026 (3 June 2026) reg 4.
[53] Press Information Bureau, ‘The Digital Transformation of the Indian Judiciary: e-Courts Mission Mode Project’ (Backgrounder, 7 August 2026) https://static.pib.gov.in/WriteReadData/specificdocs/documents/2026/aug/doc202687946901.pdf.
[54] UNESCO, Global Judges’ Initiative: Survey on the Use of AI Systems by Judicial Operators (2024); UNDP, ‘Asia-Pacific Risks Widening Justice Gaps without Action on AI, UNDP Warns’ (22 June 2026) https://www.undp.org/asia-pacific/press-releases/asia-pacific-risks-widening-justice-gaps-without-action-ai-undp-warns.

