Framework for the Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI)

A committee established by the Reserve Bank of India (RBI) has submitted a report on a framework for the Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI) in the financial sector. The report provides a blueprint for fostering AI innovation while simultaneously mitigating its risks.

Foundational Principles: The 7 Sutras

The report is anchored in seven core principles, or ‘sutras,’ which serve as the ethical foundation for all AI adoption in the financial sector. They are:

1. Trust is the Foundation: Building confidence among all stakeholders.
2. People First: Prioritizing human interests and keeping humans in the loop.
3. Innovation over Restraint: Encouraging growth without stifling progress through excessive regulation.
4. Fairness and Equity: Ensuring AI-driven decisions are free from bias and discrimination.
5. Accountability: Establishing clear responsibility for AI-driven outcomes.
6. Understandable by Design: Promoting transparency and explainability in AI systems.
7. Safety, Resilience, and Sustainability: Ensuring long-term reliability and security of AI systems.

Key Recommendations and Strategic Pillars

Guided by these principles, the committee has provided 26 actionable recommendations under six strategic pillars. The framework’s core vision is to treat innovation and risk mitigation as complementary, not competing, objectives.

Innovation Enablement Pillars:

• Infrastructure: The committee recommends creating shared infrastructure, such as a Financial Sector Data Infrastructure (FSDI), to democratize access to data and computing resources. It also proposes an AI Innovation Sandbox for regulated entities to test new AI solutions in a controlled environment.
• Policy: It recommends flexible and adaptive regulations, including a graded liability framework that offers some tolerance for first-time errors to encourage innovation. It also calls for a permanent AI Standing Committee under the RBI to provide ongoing guidance.
• Capacity: The report emphasizes the need for training programs and upskilling for boards, executives, and regulators to ensure they have the necessary expertise in AI.

Risk Mitigation Pillars:

• Governance: The committee advises that all regulated entities must have a board-approved AI policy and a clear governance structure.
• Protection: It recommends expanding consumer protection frameworks and enhancing cybersecurity measures to address AI-specific risks, such as privacy breaches and sophisticated fraud.
• Assurance: The report calls for a risk-based AI audit framework, including periodic internal and third-party reviews, and recommends that regulated entities maintain a complete inventory of all AI systems they deploy.

(Source: TH)