By Inderjit Badhwar
There are few institutions in a democracy whose credibility is as fundamental to its survival as the judiciary. Governments may change, political fortunes may rise and fall, and public opinion may fluctuate, but the authority of the courts rests on something more enduring: the belief that justice is administered through established principles, verified facts, reasoned arguments and the faithful application of law. It is this foundational confidence that makes the subject of our cover story, The Judgment That Never Was, particularly important.
Artificial Intelligence has arrived with extraordinary promises. It can retrieve information, analyse enormous volumes of material, identify patterns, summarise complex documents and assist professionals in ways that would have been unimaginable only a few years ago. The legal profession is understandably among those discovering its possibilities. But there is a fundamental distinction between a machine helping us locate the law and a machine inventing law that does not exist.
That distinction is at the heart of this investigation.
The cases examined in this issue, including those involving Deepak Raheja, Gummadi Usha Rani, Pooja Ramesh Singh and Vijay Ghanshyam Gadiya, bring into focus a troubling development: fabricated legal material and fictitious judicial precedents finding their way into the process of legal argument and scrutiny. The wider international experience, reflected in cases such as Mata vs Avianca and Ayinde, shows that this is not an isolated Indian problem or a temporary technological aberration. It is a challenge confronting legal systems across jurisdictions.
What makes the phenomenon particularly disturbing is the peculiar nature of the legal profession’s dependence on precedent. A judgment is not merely a document or an opinion. It represents an exercise of judicial authority, grounded in an actual dispute, an established legal process and the reasoning of a court. Its citation carries the weight of institutional legitimacy. A fictitious judgment, however convincingly presented, possesses none of these attributes.
Yet, generative AI can produce references, quotations, case names and legal reasoning with an appearance of remarkable authenticity. The danger is not simply that a lawyer may submit an incorrect citation. It is that a fabricated authority may enter a judicial proceeding, consume valuable court time, mislead the parties or, in the worst circumstances, influence the reasoning through which justice is delivered.
We must, therefore, ask a difficult question: who is responsible when the machine invents the law?
The answer cannot be that the machine is responsible. AI has no professional oath, no ethical obligation, no understanding of the consequences of an erroneous submission and no accountability before a court. Responsibility must remain with the human beings who deploy it. Lawyers cannot surrender their duty of verification to an algorithm. Litigants cannot assume that technological sophistication guarantees accuracy. And courts cannot afford to regard the problem as merely another procedural irregularity.
There is also a larger institutional dimension. The judicial system is already burdened by enormous pendency, limited resources and the challenge of ensuring timely access to justice. Every fabricated authority that enters a courtroom adds another layer of avoidable uncertainty. Judges must spend time checking whether a cited judgment exists, whether the quotation is genuine and whether the proposition attributed to it was ever laid down. The cost is ultimately borne by the institution and by litigants waiting for their disputes to be resolved.
But there is a danger in going too far in the opposite direction as well. Artificial Intelligence is not inherently inimical to justice. Properly used, it can improve legal research, help organise evidence, assist in translation and make legal information more accessible. The objective should not be to retreat from technology, but to establish clear boundaries around its use. Human verification, professional responsibility, transparency and appropriate judicial safeguards must accompany its adoption.
The distinction is between using technology as an instrument and allowing it to become an unexamined authority.
There is a philosophical question here too. Law is not simply a collection of searchable words, citations and propositions. It is an evolving body of human reasoning, shaped by constitutional values, competing rights, precedent, interpretation and the lived consequences of judicial decisions. Artificial Intelligence can reproduce the language of legal reasoning. It cannot, by that achievement alone, acquire the authority or responsibility that makes a judgment legitimate.
The title of our cover story is deliberately stark: The Judgment That Never Was. It describes more than an imaginary case or an invented citation. It points to a moment when the appearance of legal truth can become detached from truth itself. In a profession built on evidence and authenticity, that is an alarming possibility.
The challenge before India’s legal institutions is to recognise this danger early, respond proportionately and ensure that technological progress strengthens rather than diminishes public confidence in justice.
A court must be able to distinguish between an argument and an authority, between a plausible-looking citation and an actual judgment, between what a machine can generate and what the law has genuinely established.
Ultimately, the legitimacy of justice cannot be outsourced. Artificial Intelligence may help us search the law, understand it and navigate its complexities. But it must never be permitted to invent the law and pass it off as truth.
A judgment that never was must never become a judgment that is believed.
The post When The Law Begins To Believe Fiction appeared first on India Legal.
