AI on the docket: Justice Manmohan says automation could resolve 60% of cases
AI can clear a major share of India's pendency if it handles routine, small-ticket matters. Speaking at the India Law, AI and Tech Summit 2025 in Delhi, Supreme Court Justice Manmohan said shifting these disputes to AI-assisted decisions could resolve more than 60% of litigation and free courts to focus on the hardest cases.
His point was simple: let technology do the repetitive work; let judges apply judgment where it truly matters.
What could move to AI now
Justice Manmohan cited case types that are largely standardized and fact-light. These are the matters that clog dockets without demanding nuanced judicial reasoning.
- Traffic challans
- Cheque bounce matters
- Other small-ticket, repetitive offenses
He added that AI could cluster thousands of similar filings-like land acquisition disputes-and enable block disposals through a single ruling, cutting months of administrative drag.
Inside the Supreme Court pilot
The Supreme Court has begun piloting an AI-driven tool, described as a "digital research assistant," to read case files, extract issues, and flag relevant precedents for judges. "It prepares a summary of the readings, the issues, and calls out the law… It will not give a judgment, but it highlights what is relevant for the judge to consider," he said.
The goal is to reduce time spent sifting through bulky paper books and voluminous filings, so benches can concentrate on core adjudication. For context on institutional initiatives, see the Supreme Court of India.
Guardrails matter
Justice Manmohan was clear about the risks: hallucinated case law, algorithmic bias, and privacy concerns. These must be addressed before wide deployment.
Expect hybrid models: AI handles routine and administrative load; judges bring reasoning, empathy, and constitutional sensitivity. As he put it, the character of justice ultimately rests on the integrity, independence, and intellect of the human judge.
The counterpoint from the Bar
At the same session, Dr. Lalit Bhasin warned against over-reliance on AI. In interviews with young lawyers, he saw identical drafts-an early sign of copy-paste thinking. "There is very minimal application of mind… a dangerous course of action so far as this profession is concerned," he said.
He also highlighted that pendency is a structural problem, not just a tooling problem: more than 5 crore cases were pending in 2023-24, and the number is still rising. The deeper causes, he argued, include too many, outdated, overlapping, and ill-drafted laws. Tools help-but they cannot substitute legislative cleanup, process reform, and disciplined case management.
What legal teams can do now
- Map your docket: separate standardized matters from those needing deep judicial evaluation.
- Pilot human-in-the-loop workflows for traffic, 138 NI Act, and other routine matters.
- Standardize inputs: templates for facts, issues, and reliefs to reduce variance and speed review.
- Adopt validation layers: precedent verification, red-teaming for bias, and privacy checks.
- Track outcomes: measure time-to-disposal, error rates, and appeals to refine the model's scope.
- Invest in training: upskill teams on AI literacy, prompt discipline, and ethical use.
For principles on trustworthy AI-accountability, transparency, and human oversight-review the OECD AI Principles.
Bottom line
There's a clear near-term win: let AI take first pass on routine, high-volume matters while judges decide the cases that shape rights, markets, and public life. That split can shrink pendency without diluting justice.
But discipline is non-negotiable-strong governance, clear scope, and constant human oversight. Use the tool; don't become its echo.
Resources
For teams building practical AI skills for legal operations and research, see curated programs by role: AI courses by job.
Your membership also unlocks:
