AI Regulation Tracker / Reports and frameworks
India's Central Bank Publishes the FREE-AI Report: 7 Sutras and 26 Recommendations for AI in Finance
According to the RBI's press release, on August 13, 2025 the Reserve Bank of India released the report of its committee on a Framework for Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI) in the Financial Sector. The committee set out 7 Sutras as foundational principles and 26 actionable recommendations across six strategic pillars. This is an advisory report, not a rule. Nothing in it binds a bank or a fintech today; in my read, it is likely the blueprint the RBI's later, enforceable AI work will build on.
Here is the sequence, because it matters for how much weight to give this. In its Statement on Developmental and Regulatory Policies of December 6, 2024, released alongside the monetary policy statement, the RBI announced it would form a committee on responsible AI in finance. It constituted that committee on December 26, 2024. After what the RBI describes as extensive deliberations with diverse stakeholders, the committee submitted its report, and on August 13, 2025 the RBI placed the report on its website and issued a press release announcing it.
The framing is deliberately two-sided. The RBI says the report "sets out a framework to guide the use of AI in the financial sector, aiming to harness its potential while safeguarding against associated risks." The closing line of the press release makes the same point plainly: the report envisions a financial ecosystem where encouraging innovation is in harmony, and not at odds, with the mitigation of risk. That is the whole posture. Not a crackdown, and not a blank cheque either.
What are the 7 Sutras and the 26 recommendations?
The committee built the report on what it calls 7 Sutras. In the RBI's words, these "serve as the foundational principles for AI adoption." Guided by those principles, the committee then recommended what it describes as a forward-looking approach, "containing 26 actionable recommendations under six strategic pillars."
The structure is worth understanding even before the detail. Seven principles at the top set the values. Six pillars underneath organize the work into strategic areas. Twenty-six concrete recommendations sit inside those pillars as the things a regulator or a firm could actually do. That shape, principles then pillars then actions, is how a report like this becomes a roadmap rather than a wish list. The recommendations are addressed to the ecosystem, which includes the RBI itself, regulated entities, and the wider market, not to any single firm as an obligation.
Is this a rule banks have to follow?
No, and this is the part to get right. A committee report is not a regulation. The RBI convened experts, took their recommendations, and published the result for guidance and public discussion. Nothing in the report requires a bank, an NBFC, or a fintech to change a control, file anything, or meet a deadline. There is no penalty for non-alignment, because there is nothing yet to be out of compliance with.
That distinction is easy to lose when a central bank publishes a framework, because the source carries authority. But authority to convene and recommend is not the same as a direction that binds. If and when the RBI wants any of these 26 recommendations to have force, it has to issue its own instrument, a circular or a master direction, through its normal process. Until it does, FREE-AI is a signal about where the regulator's thinking is heading, not a live requirement.
How this differs from the binding RBI AI drafts
This is where people get tangled, so I want to separate two things that are not the same document. FREE-AI, the report covered here, is advisory. It is the high-level framework. Separately, the RBI has been moving on binding model-risk material, including a draft touching model governance and controls of the kind that would actually constrain how regulated entities deploy models. That binding track is a different instrument on a different timeline, and this tracker follows it on its own page.
Keep them apart in your own notes. When someone says "the RBI's AI rules," ask which one. The FREE-AI report is the principles-and-recommendations layer with no legal force. The model-risk draft is the enforceable layer. Conflating the two leads to either overreacting to a report or underreacting to a rule, and both are expensive mistakes.
Why a US professional should care
Two reasons, neither of them "India passed a law," because it did not. First, if you run or advise financial operations with exposure to India, a subsidiary, a lending partner, a payments arrangement, an outsourced analytics function, this report is the clearest available preview of the compliance regime those operations will eventually sit under. The recommendations tell you what the RBI's own experts think good AI governance in finance looks like. Building toward that now is cheaper than retrofitting after a binding direction lands.
Second, even with no Indian footprint, FREE-AI is a useful comparative benchmark. It is a major central bank in a large market laying out how it wants AI governed in banking, and the principles-pillars-actions structure is portable. For a CPA, a consultant, or an executive assembling an internal AI policy, a published regulator framework is a credible reference point you can borrow from, provided you remember it is Indian in origin and advisory in status. Use the judgment, not the jurisdiction.
Questions professionals are asking
Is the FREE-AI report a binding regulation?
No. It is a committee report the RBI commissioned and published for guidance on August 13, 2025. It sets out principles and recommendations, but it creates no legal obligation and imposes no deadline or penalty. Any binding effect would require the RBI to issue a separate instrument such as a circular or master direction.
What are the 7 Sutras and 26 recommendations?
The committee developed 7 Sutras, which the RBI describes as the foundational principles for AI adoption. Guided by those principles, it made 26 actionable recommendations organized under six strategic pillars, aimed at encouraging innovation while mitigating the risks of AI in the financial sector.
How is this different from the RBI's binding AI draft?
FREE-AI is the advisory framework, with no legal force. The RBI is separately developing binding material on model risk and governance that would actually constrain how regulated entities use models. They are different documents on different timelines. Treat the report as the blueprint and the model-risk draft as the enforceable rule.
Does this affect a US financial firm?
Not as law. If your operations touch India through a subsidiary, partner, or outsourced function, the report is the clearest preview of the AI governance regime those operations will eventually face, so it is worth building toward. With no Indian footprint, it is still a useful comparative benchmark for internal AI policy, as long as you treat it as advisory and Indian in origin.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any framework, report, or requirement applies to your situation with qualified professionals in the relevant jurisdiction.