The FREE-AI Framework: A New Era for Ethical AI in Indian Finance

I. Introduction: The Rise of AI and the Need for a Guiding Hand

In response to the rapid and transformative rise of artificial intelligence (AI), the Reserve Bank of India (RBI) set up a high-level committee in December 2024 to develop a guiding framework. After months of extensive consultations and two nationwide surveys of financial institutions, the committee released its landmark report, establishing the "Framework for Responsible and Ethical Enablement of Artificial Intelligence" (FREE-AI). The FREE-AI serves as a blueprint for how banks and financial services will build and deploy AI systems in India.

The FREE-AI provides a comprehensive blueprint for ethical and responsible AI adoption in India’s financial sector, balancing innovation with robust risk mitigation. Anchored in seven guiding principles and 26 actionable recommendations across six pillars, it establishes standards for governance, capacity building, consumer protection, and compliance. It aims to drive inclusion, transparency, and trust while positioning India as a global leader in responsible AI-driven finance.

II. The "Seven Sutras": Guiding Principles of the FREE-AI

At the heart of the FREE-AI lie seven guiding principles. These principles serve as the ethical and philosophical bedrock for all AI development and deployment in the financial sector and are listed as follows:

1. Trust is the Foundation: All AI systems must be reliable, transparent, and designed to inspire public confidence. Without these qualities, adoption slows, trust erodes, and the technology faces greater regulatory and reputational risk.
2. People First: AI must serve as a tool to augment human decision-making, not replace it. This principle ensures that AI is used to improve human welfare and dignity, with a clear deferral to human judgment and citizen interests.
3. Innovation over Restraint: The purposeful development of AI should be prioritised, consciously avoiding unnecessary restrictions that could stifle technological progress and economic growth.
4. Fairness and Equity: AI outcomes must be fair and non-discriminatory, preventing models from inheriting and perpetuating biases from their training data. If unchecked, these biases can reinforce inequality, undermine trust, and expose organizations to legal and ethical risks.
5. Accountability: Accountability for the outcomes of an AI system rests solely with the entity that deploys it, regardless of the system's level of autonomy. This principle clearly assigns responsibility for AI decisions and their impacts.
6. Understandable by Design: AI systems must be interpretable by design. This principle requires that the logic behind an AI's decision can be traced and understood by both end-users and regulators, moving away from opaque "black box" models.
7. Safety, Resilience, and Sustainability: AI systems to be secure, resilient to shocks, adaptable to changing conditions, and energy-efficient for long-term viability. This principle follows a forward-looking approach.

III. The "Six Pillars": A Strategic Approach to Implementation

To translate these guiding principles into actionable directives, the FREE-AI outlines 26 recommendations organized under six distinct pillars. These pillars are deliberately divided into two complementary categories:

1. Innovation Enablement.
2. Risk Mitigation.

1. Innovation Enablement

This category focuses on creating an environment where responsible AI innovation can thrive.

Infrastructure: To support the infrastructure, it recommends the following:

1. Build a high-quality financial sector data infrastructure as a digital public good, integrated with the India AI Mission’s AI Kosh platform to support trustworthy AI models.
2. Establish an AI Innovation Sandbox for the financial sector, allowing Regulated Entities (REs), fintechs, and innovators to safely develop and test AI-driven solutions, with collaboration from other FSRs.
3. Provide incentives and funding to support smaller entities, promoting inclusive AI usage and enabling the creation of data and compute infrastructure for sectoral innovation.
4. Encourage the development of indigenous financial sector AI models, including LLMs, SLMs, and other specialized models, as a public good.
5. Establish a framework to integrate AI with Digital Public Infrastructure (DPI) to accelerate inclusive and affordable financial services at scale.

Policy: For policy enablement, it recommends the following:

1. Implement adaptive policies by periodically reviewing existing frameworks, developing a comprehensive AI policy for the financial sector based on the Committee’s 7 Sutras, and issuing consolidated AI guidance for responsible innovation and risk management.
2. Promote AI-driven initiatives that enhance financial inclusion for underserved populations by easing compliance requirements where possible, while maintaining essential safeguards.
3. Adopt a graded AI liability framework that balances responsible innovation with accountability, allowing regulatory tolerance for first-time issues when safety mechanisms are followed, while maintaining full liability for repeated breaches or gross negligence.
4. Establish a permanent multi-stakeholder AI Standing Committee under the RBI to advise on emerging risks and opportunities, and create a dedicated financial sector AI institution linked to the national AI Safety Institute for continuous monitoring and coordination.

Capacity: For capacity building, it suggests the following:

1. Build AI capacity within REs through board- and C-suite-level governance, and provide ongoing training and upskilling for staff to ensure responsible and ethical AI adoption.
2. Strengthen regulator and supervisor capacity through targeted AI training and institutional initiatives, including a dedicated AI institute, to keep frameworks aligned with evolving technologies and associated risks.
3. Create a framework for sharing AI best practices across the financial sector, enabling the exchange of use cases, lessons learned, and governance insights to promote responsible scaling.
4. Introduce programs to recognize and reward responsible AI innovation in the financial sector, emphasizing ethical design and positive social impact.

2. Risk Mitigation

This category outlines the institutional and operational safeguards necessary for safe AI deployment.

Governance: For improved governance, it suggests the following:

1. Require REs to adopt a board-approved AI policy covering governance, accountability, risk, safeguards, auditability, consumer protection, and model lifecycle, with industry guidance for smaller entities.
2. Implement robust data lifecycle governance in REs, covering collection, access, usage, retention, and deletion, ensuring compliance with laws like the DPDP Act.
3. Establish comprehensive AI system governance in REs, covering the full model lifecycle, including design, deployment, monitoring, and decommissioning, with human oversight for autonomous and high-risk applications. OR Establish comprehensive AI system governance in REs.
4. Include all AI-enabled products in REs’ product approval frameworks, with mandatory AI-specific risk evaluations.

Protection: For enhanced protection, it suggests the following:

1. Implement a board-approved consumer protection framework in REs, ensuring transparency, fairness, accessible recourse, and ongoing consumer education on safe AI use.
2. Strengthen AI-related cybersecurity in REs by identifying risks, enhancing infrastructure and processes, and leveraging AI tools for threat detection and response.
3. Implement structured red teaming across the AI lifecycle in REs, with frequency and intensity aligned to risk levels and triggered by evolving threats or changes.
4. Enhance REs’ business continuity plans to cover AI system failures and model performance issues, with fallback mechanisms and regular resilience testing.
5. Establish an AI incident reporting framework for REs and fintechs, promoting timely detection and good-faith reporting of AI-related issues.

Assurance: To ensure oversight throughout the AI lifecycle, it recommends the following:

1. Maintain a comprehensive AI inventory within REs, updated biannually for supervisory review, and establish a sector-wide repository to track adoption trends, risks, and systemic vulnerabilities.
2. Implement a risk-based AI audit framework in REs, including internal audits, independent third-party audits for high-risk cases, and biennial reviews, with supervisors providing clear guidance on audit standards and compliance.
3. Require REs to provide AI-related disclosures in annual reports and websites, following a regulator-specified framework for consistency and transparency.
4. Develop an AI Compliance Toolkit, maintained by a recognized SRO or industry body, to help REs validate and demonstrate adherence to key responsible AI principles like fairness, transparency, accountability, and robustness.

IV. Opportunities and Challenges Ahead

• Purpose & Opportunity

1. 1. Not just defensive—enables the next phase of digital finance in India.
   2. Provides a clear, trustworthy regulatory path for responsible AI use.
   3. Drives financial inclusion by using alternative data to assess creditworthiness for unbanked and underbanked populations.
   4. Enhances efficiency and risk management, e.g., AI can reduce fraud and lower account rejection rates by 15–20%.
   5. Lowers barriers for fintech startups, encouraging innovation and investment.

• Implementation Challenges

1. 1. Cultural shift is the main hurdle: banks must embed fairness and transparency from day one.
   2. Data quality: Legacy data needs cleanup before AI deployment.
   3. Governance & workforce: Requires new governance structures and upskilling across staff.
   4. Auditing complexity: Compliance is not just algorithm testing; it includes governance processes, use case approvals, and documentation.
   5. Smaller institutions may feel pressure due to upfront investments.

• Strategic Impact

1. 1. FREE-AI elevates AI adoption from “best practice” to mandatory compliance, setting a new industry standard.
   2. Successful implementation positions banks to leverage AI responsibly, improve inclusion, and foster trust with regulators and customers.

V. What this Means for the Future of Finance

The FREE-AI’s vision is a financial ecosystem where encouraging innovation is in harmony, and not at odds, with the mitigation of risk. If implemented, this comprehensive approach could serve as a global model for how other emerging economies can responsibly harness the transformative potential of AI while preserving public trust and financial integrity.

The path forward is clear: the FREE-AI positions India as a potential leader in ethical, inclusive, and safe AI adoption, ensuring that the next era of digital finance is built on a foundation of trust for all citizens.

VI. How Securiti Supports Compliance with the FREE-AI Framework

Securiti’s AI Security & Governance solution, part of Securiti Data Command Center, delivers on several core dimensions that are highly relevant to FREE-AI’s requirements. By delivering comprehensive AI visibility, risk-based compliance, governance structures, and data controls, it supports:

• Ethical and explainable AI,
• Robust governance and oversight,
• Continuous protection and compliance,
• Building organizational capacity for responsible AI adoption.

Request a demo to learn more.