Nasscom urged the Supreme Court to define "high-risk AI" and clarify audit procedures in its draft regulations for court use. The industry body fears vague terms could lead to inconsistent interpretations and mandatory source code disclosures for third-party vendors. This creates a direct tension between judicial oversight and AI vendor IP protection.
The Supreme Court’s AI Committee, led by Justice P.S. Narasimha, released its draft AI regulations for courts on June 3, inviting comments until June 20. These rules aim to govern AI use across all Indian courts, permitting legal research and transcription but specifically banning AI for adjudication or sentencing.
The Supreme Court's AI Committee is now reviewing all public submissions received by the June 20 deadline. Expect a revised draft or final version of these regulations within the next 60-90 days, indicating how Nasscom's concerns were addressed.
🇮🇳 Why This Matters for India
For legal tech founders in Bangalore and product managers building AI tools for government clients, these definitions dictate what they can build and how much IP they might need to expose.
The Take
Mandating white-box audits for all AI vendors would force courts into expensive, custom-built solutions, killing the adoption of off-the-shelf, low-risk tools. The smarter play is adopting risk-based API-level audits for most deployments and only requesting deeper access when actual risk warrants it.
Source:  MediaNama ↗