Reserve Bank of India Proposes AI Kill Switch Rules

The Reserve Bank of India released a draft framework titled Guidance on Regulatory Principles for Model Risk Management, 2026, to regulate the use of artificial intelligence and machine learning in the financial sector. The proposed rules apply to commercial banks, co-operative banks, payment banks, and non-banking financial companies (NBFCs).

A central requirement of the framework is the implementation of a kill switch to instantly override, suspend, or deactivate AI models that produce harmful or erroneous outputs. To prevent automation bias and ensure accountability, the guidelines mandate human oversight for AI-driven decisions and require customer-facing systems to disclose their AI identity and provide an option to switch to human support.

Regulated entities must establish board-approved risk management policies, maintain model inventories, and utilize a risk-based tiering structure. High-risk models will require specific approval from the Board's Risk Management Committee and must undergo rigorous validation and stress testing. The bank also highlighted the need for cybersecurity protections against prompt injection attacks.

The framework asserts that institutions remain fully accountable for model outcomes, including those sourced from third-party vendors, to mitigate supply chain risks associated with dominant AI providers. The central bank has invited public feedback on the draft guidelines through July 24, 2026.