Implementation of RBI's FREE-AI Framework
The Reserve Bank of India’s (RBI) new FREE-AI framework emphasizes fairness, robustness, efficiency, and explainability in AI systems to enhance trust and accountability within the banking and financial services sectors.
Key Requirements for Implementation
- Significant investment in infrastructure and training.
- New governance structures.
- Sector-wide capacity building.
- Cultural change to instill fairness and transparency from the outset.
Expert Opinions
Ajay Trehan, CEO of AuthBridge, highlights the importance of cultural change and define key components:
- Explainability: AI decisions must be traceable to human-understandable logic.
- Fairness: Ensuring no demographic is unfairly disadvantaged.
- Long-term benefits: Include improved decision quality, reduced compliance risk, and enhanced customer trust.
Karthik Pasupathy from EY India emphasizes broader evaluation strategies:
- Auditing: Involves evaluating governance processes, use case approvals, and risk and compliance functions.
- Models need to be tested against rare or extreme scenarios to ensure compliance with laws and ethics.
- Techniques like SHAP and LIME can elucidate complex models, but fairness hinges on quality training data.
- Data cleaning: The challenge lies in removing historical biases from datasets.
Most Indian banks are currently in the early stages of AI adoption, transitioning from rule-based systems to simpler machine learning models.