AI-based resolution of cheque bounce cases is feasible: Ex CJI DY Chandrachud
Ex Chief Justice of India Justice DY Chandrachud signalled a pragmatic path for AI in Indian courts: start with high-volume, transactional disputes like cheque dishonour matters. His point was straightforward-use AI where outcomes are largely standardised and rights impacts are limited, keep humans where stakes are high or facts are nuanced.
He delivered these remarks during the IBA Litigation and ADR Symposium session on the benefits and impact of AI on dispute resolution, tying technology choices to constitutional values: efficiency, access, and fairness.
Why cheque bounce cases fit AI
Cheque dishonour litigation under Section 138 has flooded magistrate dockets for years. The issues are repeat, the statutory ingredients are consistent, and the reliefs are predictable. That makes them prime candidates for AI-enabled adjudication with clear templates and explainable outputs.
Justice Chandrachud noted India's experience with Delhi's virtual courts for routine traffic offences-an automation step that freed up dozens of magistrates to hear matters requiring real judicial attention. A similar model could streamline cheque-bounce cases, reduce pendency, and preserve time for complex disputes. eCourts Virtual Courts offer a live reference point for how such systems can be structured.
Where humans must remain in the loop
He drew a clear line on areas unsuitable for automated disposition. Housing and rent control matters risk eviction and displacement-human oversight is essential. Motor accident compensation could adopt an optional AI route: instant, explainable awards that bind insurers while giving victims a choice to accept or seek judicial adjudication.
Efficiency as a constitutional value
Justice Chandrachud highlighted the tension between exhaustive procedural safeguards and timely relief. "Efficiency must itself also be a constitutional value," he said, warning that without it, "you are not giving timely justice to those who need it the most."
He added a hard truth for system design: "At some stage, we have to balance the two and allow for some degree of injustice within the system, so as to make the system overall efficient," cautioning that trying to correct every error can "render every court dysfunctional."
Implications for legal teams
- Criminal complaints under NI Act: Expect standardised scrutiny of statutory ingredients, timelines, and presumptions. Documentation discipline will matter more than rhetoric.
- Insurers and TPAs: Prepare for optional AI awards in motor claims; align policy wording, data pipelines, and escalation paths for claimant choice and audits.
- Prosecution and defence: For AI-routed matters, focus on data completeness, e-evidence formats, and structured submissions that map to statutory checklists.
- Judiciary administration: Reallocate judicial time to fact-heavy or rights-sensitive dockets; use dashboards to track pendency, variance, and error rates.
- In-house counsel (banking/fintech): Standardise demand notices, proof of service, and account statements. Build batch-ready, machine-readable case files.
- Ethics and due process: Push for explainability, contestation windows, accessible appeals, and bias testing. No black-box adjudication.
Guardrails for an AI adjudication pilot
- Eligibility filters: Limit to uncontested facts or narrow legal issues (e.g., cheque issuance, presentation, notice, limitation).
- Consent and choice: Clear opt-in for parties where appropriate; guaranteed path to human review on request or defined thresholds.
- Explainability: Structured reasons referencing statute, evidence, and presumptions; machine-generated orders must be intelligible.
- Appeals and oversight: Fast-track human review with defined timelines; audit trails for every automated step.
- Data governance: Secure evidence ingestion, standard formats (PDF/A, machine-readable metadata), retention and privacy controls.
- Fairness controls: Bias testing, error-rate monitoring, and regular independent audits; publish aggregate performance metrics.
- Sandbox approach: Start with limited courts and case types; iterate based on variance, reversal rates, and user feedback.
Impact on the profession
Justice Chandrachud described AI as a disruptor that will shift effort from repetitive tasks to higher-order advocacy. Some traditional functions may shrink, but well-designed systems can expand access and lighten judicial burden-if dignity, contestability, and transparency are built in from day one.
The message to the bar is clear: adapt workflows, sharpen statutory precision, and learn to work with explainable tools. Technology should accelerate justice, not replace judgment.
Session context
These remarks were delivered in a keynote on technology, constitutionalism, and the future of dispute resolution at the IBA Litigation and ADR Symposium. The session was chaired by Carlo Portatadino (Tombari D'Angelo e Associati, Milan) with a panel featuring Jayant Mehta (Senior Advocate, New Delhi), Mahesh Rai (Drew & Napier, Singapore), and Professor Tania Sourdin (Professor Emerita, University of Newcastle, Australia). For background on the forum, see the IBA Litigation Committee.
Next steps for legal teams
- Map your Section 138 workflow end-to-end; convert recurring tasks into templates and structured data.
- Pilot explainable AI tools internally for intake, document checks, and statutory compliance before any court-facing automation.
- Advocate for safeguards: clear reasons, human review, claimant choice in compensation matters, and published performance data.
If you're upskilling teams on practical AI for legal operations, explore role-based learning paths here: Complete AI Training - Courses by Job.
Your membership also unlocks:
