The New Rulebook: RBI’s Framework (FREE-AI) for Responsible AI in Finance
From privacy-first to trust-first: how India’s FREE-AI framework is reshaping the future of financial innovation.
This week on Eximius Echo, we’re tracking a shift at the heart of India’s financial system: the move from data protection to responsible AI adoption.
The DPDP Act secured citizens’ right to privacy. Now, with AI reshaping everything from credit underwriting to fraud detection, the RBI has introduced the FREE-AI framework, a set of guiding principles to ensure innovation doesn’t come at the cost of trust.
Until now, adoption was slow - just ~1 in 5 regulated entities had begun experimenting with AI, held back by compliance concerns and the “black box” problem. FREE-AI signals a new era: one where banks, NBFCs, and fintechs can scale AI confidently, backed by clear guardrails.
For founders, this unlocks massive white spaces: indigenous language models built on DPIs, AI-first RegTech platforms, cybersecurity layers against deepfakes and adversarial threats, and inclusive credit models for underserved segments.
The message is clear, AI in finance is no longer optional, but its future will be defined not by speed alone, but by trust, resilience, and accountability.If you’re new here, Eximius is a Pre-seed VC fund backing bold ideas in Fintech, AI/SaaS, and Consumer Tech. We use this newsletter to share insights, trends, and ideas from the sectors we’re passionate about. Let’s dive in.
India took a landmark step with the Digital Personal Data Protection (DPDP) Act, establishing a strong foundation for citizens’ right to privacy in the digital age. But as Artificial Intelligence rapidly reshapes finance and FinOps, privacy alone is no longer enough. Questions around accountability, governance, and policy - for citizens, financial institutions, and AI-first startups - have come to the forefront. Without clear guardrails, AI risks amplifying bias, eroding trust, discriminating against users, and creating new systemic vulnerabilities.
Recognising this dual reality, the RBI introduced the Framework for Responsible and Ethical Enablement of AI (FREE-AI) - not as an act, but as a guiding framework to improve adoption and bolster confidence in the AI ecosystem.
Through this framework, the RBI’s objective is to establish a balanced, forward-looking approach that not only fosters innovation but also builds and sustains public trust.
In this newsletter, we highlight key areas of impact from the framework: how banks and regulated entities can work within it, and how founders building at the intersection of AI and Fintech can leverage AI responsibly to cater to the needs of banks and NBFCs.
The Current State of AI Adoption by Fintechs
AI has really taken off over the last couple of years, especially with tools like ChatGPT making it easier for anyone to tap into AI for research, writing, and personal projects. But when you look at India’s regulated entities (REs), AIFs, Asset Reconstruction Companies, Banks, and NBFCs, adoption is still in its early days. Right now, only about 21% of all REs are either using AI or actively developing AI systems. Among NBFCs specifically, just 27% have started leveraging AI to run and optimize their processes.
Nevertheless, this pace of adoption is actually faster than what we’ve seen with most other emerging technologies in such a short time frame. Banks and NBFCs are already utilizing AI for high-impact use cases from customer support chatbots and smarter credit underwriting, to automating sales and marketing workflows, and even strengthening their cybersecurity frameworks.
Still, despite these early signs of progress, there’s a noticeable sense of hesitation when it comes to scaling AI across the board. Many financial institutions are wary of fully trusting AI because of its “black box” nature, decisions can feel unpredictable and hard to explain.
To build confidence and unlock large-scale adoption, what’s really needed is a clear, broad governance framework, one that gives both regulated financial institutions and fintech founders the guiding principles they need to deploy AI responsibly and at scale.
The FREE-AI Framework
The FREE-AI framework establishes the RBI’s proactive approach to sustainably and responsibly adopting AI. To represent this, the framework highlights key principles that any regulated entity or founder should adopt to leverage AI to the fullest:
First, it must be people-first - AI should augment human decision-making and enable better outcomes.
Second, innovations in AI must be developed responsibly and aligned with societal values, with the aim of maximizing inclusivity and benefits while minimizing potential harm. With responsible AI innovation comes accountability: entities deploying AI, not the models themselves, must be held fully accountable for decisions and outcomes.
Finally, these principles are tied together with the idea of safety, resilience, and sustainability - AI systems must be secure, resilient to physical, infrastructural, and cyber risks, and designed to be energy-efficient.
By setting out these guiding principles, the RBI is encouraging Indian institutions and innovators - particularly within the IndiaAI Mission ecosystem - to step forward and build for India. These principles are highly instrumental in enabling this shift, and at Eximius, we believe they lay the foundation for how India can truly innovate with AI.
But how can we innovate in AI for India?
To truly foster AI innovation, something we at Eximius have been closely tracking is what the RBI has been encouraging: leveraging data and India’s Digital Public Infrastructure (DPIs) to build homegrown, indigenous Small and Large Language Models. This would allow regulated entities to develop their own proprietary models with deeper Indian context, giving them more control, better alignment with compliance requirements, and a real competitive edge.
With the advent of Account Aggregators and India’s open data ecosystem, the country has made huge strides in making public financial data available and securely stored on data-blind infrastructure. The next big opportunity lies in responsibly layering AI on top of this foundation.
We then need serious tech innovation in terms of robust financial sector data infrastructure, dedicated AI innovation sandboxes, and seamless integration of AI into core financial systems.
But this push for innovation must go hand in hand with serious risk mitigation. Consumers need to be protected through transparency and fairness at every step. At the same time, REs and fintech founders must focus on building robust systems that can detect and respond to new cybersecurity threats introduced by AI including risks like data poisoning and adversarial attacks.
This opens up quite a few whitespaces that we at Eximius are very excited about:
Custom Small Language Models (SLMs): Narrowly focused and indigenous SLMs with Indian context built on top of the Account Aggregator Layer and DPI Infra for specific financial-sector tasks enabling faster training, lower compute costs, and higher efficiency.
Agent Observability & Human-in-the-Loop Systems: Monitoring layers that oversee autonomous AI agents, ensuring compliance with policies and preventing risks like algorithmic collusion
AI Cybersecurity Stack: Cybersecurity tech that oversees threats like data poisoning, adversarial prompts, and deepfake-enabled KYC fraud that prevent model hallucinations.
AI + RegTech Platforms: AI-first compliance platforms that automatically flag, log, and reconcile breaches with relevant regulations helping compliance teams update policies dynamically and maintain governance standards.
AI-Powered Alternate Credit Models: Inclusive Credit Models that use alternate data to design credit-scoring models that expand credit access for underserved segments.
Conclusion
AI has grown leaps and bounds in the last couple of years, especially since the launch of ChatGPT. Institutions large and small are experimenting with deploying AI across diverse use cases. With AI proliferating across both enterprise and consumer contexts, it was inevitable that innovation would require a guiding playbook - one that anchors decision-making, responsibility, and governance. The RBI, being forward-looking, has become one of the first central banks in the world to publish such a framework.
With the framework, the RBI is providing a clear playbook for both fintech founders and financial institutions.
For founders, this means building products that balance innovation with compliance and risk management to increase the odds of pilots scaling into production. For financial institutions, it highlights the need to strengthen policy, governance, and accountability frameworks, creating the right environment for safe and rapid AI experimentation.
The focus on open data policies and shared infrastructure also benefits both sides, enabling founders to build more inclusive, domain-specific models while giving institutions confidence that adoption will not compromise security or compliance.
India is clearly bullish on AI, and with FREE-AI, the RBI is signaling both confidence and clarity. For financial institutions, it offers reassurance and direction on how to adopt AI responsibly. For founders, it provides a list of whitespaces to build and a clear roadmap for building sustainable, compliant solutions along with potential whitespaces to build tailored for India’s fintech ecosystem. Above all, this framework is the RBI’s way of affirming that AI is here to stay and will be a cornerstone of the future of finance in India.
If you are looking to build in this space, we would love to chat! Please reach out to us at pitches@eximiusvc.com.
 | A guest post by
|