Only 20.8% of surveyed entities using or developing AI, with adoption skewed towards large banks.
In its most comprehensive policy blueprint yet for use of artificial intelligence (AI) in finance, the Reserve Bank of India (RBI) released a ‘Framework for responsible and ethical enablement of AI’ (FREE‑AI), with a regulatory roadmap comprising 26 recommendations.
The document aims to harness AI’s transformative potential while safeguarding trust, ethics, and stability in India’s financial system.
FREE-AI was prepared by a committee set up by RBI in December 2024 and chaired by IIT Bombay professor Pushpak Bhattacharyya.
It examined global regulations, conducted industry surveys covering over 600 supervised entities, and engaged with multiple stakeholders before finalising the principles, as per an official release.
A survey conducted by the committee showed low but growing adoption of AI in the financial sector. Only 20.8% of the 612 entities surveyed by the committee were using or developing AI, with adoption skewed towards large banks.
Most deployments were in low‑risk areas like customer support, marketing, credit underwriting and cybersecurity tools. Smaller urban cooperative banks and NBFCs cited cost, data quality, and talent constraints as barriers.
The committee emphasised that trust is non-negotiable and should remain uncompromised. It highlighted the people-first approach, adding that AI should augment human decision-making, but defer to human judgment and citizen interest.
It called for fostering responsible innovation with purpose, fairness, and equity. The committee also underlined that AI outcomes should be fair and non-discriminatory. Besides, it stressed on accountability, noting that responsibility rests with the entities deploying AI. It said that AI should be understandable by design to ensure explainability for trust.
Finally, the committee stressed safety, resilience, and sustainability, stating that AI systems should be secure, resilient, and energy-efficient.
Most crucial among the 26 recommendations in the panel’s regulatory roadmap are the creation of a high-quality financial sector data infrastructure integrated with AI Kosh to democratise access to trustworthy AI models.
The committee also recommended setting up an AI innovation sandbox to allow safe experimentation in developing financial AI solutions; providing incentives, funding support, and shared compute “landing zones” so that smaller and underserved financial institutions can adopt AI.
It called for the development of indigenous, domain-specific AI models for the Indian financial sector, especially multilingual and inclusion-focused ones.
The committee also suggested adopting a graded AI liability framework that continues to hold institutions accountable but offers supervisory leniency for first-time failures with safeguards, besides embedding robust data lifecycle governance and AI model lifecycle oversight, particularly for autonomous AI systems.
Read more on Analytics India Magazine
