“We have seen our morphed pics too,” said former Chief Justice of India BR Gavai last year. It was a humorous yet striking admission: Even judges are no longer insulated from the distortions of the digital age.
With Artificial Intelligence (AI) already embedded in legal systems across the world, AI in courts is no longer a speculative question. It is used for research, translation, case management, transcription, and drafting assistance. The Supreme Court recently launched Su Sahay, an AI chatbot linked to the “One Case One Data” initiative aimed at streamlining judicial information. The question now is how deeply AI will enter the judicial system.
The case for AI is compelling. Indian courts are overburdened to the point of dysfunction. AI can organise documents swiftly and save time on administrative tasks. But it remains fundamentally a predictive technology, built on identifying patterns in data. Judicial work, however, often requires courts to break with patterns — to protect minorities against majorities, individuals against the State, and rights against convention. Judges are constantly balancing competing human realities of liberty and security, equality and tradition, punishment and mercy. AI can efficiently handle the drudgery of administration, but not judicial deliberation.
The real danger is not that robots will suddenly replace judges. It is in the algorithmic shaping of a profoundly human endeavour that is judicial reasoning. A good judgment does more than state a conclusion; it constructs a chain of reasoning that situates legal doctrine within social reality. Citizens accept difficult judgments not only for the court’s authority, but also for their coherence. Landmark judgments are valued for outcomes and the intellectual labour of balancing principles, interpreting precedents, and articulating moral responsibility.
Generative AI systems are indifferent to this craft. Their strength lies in producing plausible and standardised text quickly. As lawyers increasingly rely on AI-generated submissions and judges on AI-assisted summaries or draft structures, legal writing may become flatter and less reflective. Once AI systems begin structuring how judges receive, summarise, and prioritise information, they can indirectly shape judicial reasoning itself. The danger may not be incorrect judgments but thinner reasoning.
This is not the only concern. There is also the problem of hallucination. AI systems fabricate judgments and legal citations with alarming confidence. Once such false material enters legal databases, it can undermine the integrity of the system. Ironically, a technology expected to reduce judicial workload has instead increased it by shifting the burden towards verification.
Privacy implications further complicate the issue. Court records contain sensitive personal information. Feeding such material into AI systems inevitably raises constitutional questions about privacy, now a fundamental right.
Today, Indian courts occupy an unusual, paradoxical position. They are being asked to regulate AI, adjudicate on deepfakes, misinformation and data rights, while simultaneously integrating AI into their own institutional functioning. The judiciary is thus both regulator and consumer of the same technology.
That paradox may eventually produce difficult institutional questions. Litigants may begin questioning the role AI systems play and whether it influences judicial reasoning. In that sense, courts may face the same demands for algorithmic accountability and transparency that are demanded from other players.
The promise is that AI will eventually surpass human intelligence not only in speed and memory, but also in empathy, ethics, and creativity. Perhaps that day will come. But at present, AI does not experience human consequences. Justice, however, is built upon consequence, through judgments that alter citizens’ lives. Judicial legitimacy rests not only on correctness, but on accountability. A judge can explain reasoning, justify interpretation, and be publicly criticised. Judgments can be overturned when reasoning no longer holds. The authority of the Court is inseparable from this responsibility. None of this means AI should be rejected by the judiciary. Its integration is inevitable and, in many respects, desirable. But it must remain strictly assistive, easing administrative burden, not drafting judicial reasoning. The burden of moral and constitutional responsibility must remain firmly human.
This article was written by humans. AI may have written it faster, but speed is not the highest virtue of justice.
Insiyah Vahanvaty is a socio-political commentator and Author of ‘The Fearless Judge’. Ashish Bharadwaj is Pro Vice-Chancellor of the upcoming WPU Goa Campus. The views expressed are personal
