Nasscom seeks changes to Supreme Court's draft AI rules
India's technology association Nasscom has asked the Supreme Court for clarity on draft rules for AI use, specifically concerning the definition of high-risk applications and source code disclosure requirements for technical audits.
Nasscom, an Indian industry body, has submitted recommendations to the Supreme Court on its draft regulations for the use of Artificial Intelligence (AI) in courts. Key requests include a clearer definition of high-risk AI applications and assurance that technical audits will not automatically necessitate source code disclosure.
The Supreme Court's AI Committee released the draft earlier in June and invited comments. The proposed rules would govern AI use across all court levels for legal research, transcription, translation, and case management. However, decision-making, sentencing, and credibility assessments will remain exclusive to judges.
Nasscom highlighted that the draft fails to define terms such as "substantial risk of harm" and "high-risk applications." The organization stressed that clearer definitions and illustrative examples of risk categorization would enable courts to assess and deploy AI systems more consistently.
Furthermore, Nasscom voiced concerns regarding the regulation on technical audits, fearing it could imply that courts would gain access to AI systems' source code. This is problematic, particularly when courts utilize third-party AI services whose source code is typically not revealed. Nasscom proposes that audits should focus on system reliability, security, and explainability, with deeper code inspection reserved for high-risk or custom-built systems.
These proposed amendments reflect a broader discussion on the risks and benefits of AI regulation within the judicial system, with various concerns raised by both industry stakeholders and legal professional bodies.