Why India’s courts may have just handed legal tech its defining moment
India's courts are grappling with AI-generated fake precedents, establishing a new standard for 'defensible AI' in legal practice.Summary
In August 2025, an Andhra Pradesh trial court judge cited four Supreme Court judgments that did not exist, a case that escalated to the Supreme Court and was labeled judicial misconduct. This incident, alongside other instances of AI hallucinations in legal documents, highlights a critical vulnerability in legal AI systems: the ability to generate authoritative-sounding but non-existent citations. The Indian judiciary has historically used AI tools like SUPACE and SUVAS for assistance, but these systems are designed with human oversight to ensure verifiability. The article argues that the future of legal AI lies in 'defensible AI,' which prioritizes explainable, verifiable, and traceable outputs. This concept is gaining traction in India's regulatory framework, with guidelines emphasizing a 'techno-legal' approach that embeds accountability into system architecture from the start. As India faces a massive backlog of cases, the adoption of defensible AI is essential to maintain institutional trust and ensure that AI tools enhance, rather than undermine, the integrity of the legal system.
(Source：Cio&leader)
