I had a client call me in January this year to say she had read the MeitY Guidelines three times and still could not tell what the Sutras actually required her to do. The language was elegant. The operational meaning was elusive. I asked her whether she had read the RBI FREE-AI Report. She said no, because her organisation was not a bank. I told her to put the Guidelines down and read FREE-AI first.
She called back two days later and said she now understood the Guidelines. The confusion had not been in the Guidelines; the confusion had been in reading them in isolation. The Guidelines borrow the Sutras from FREE-AI, which spells them out in the context of a working financial-sector programme with 26 specific recommendations. Read FREE-AI first, read the Guidelines second, and the operational meaning falls into place.
The stack has two documents, in this order
The first document is the Reserve Bank of India\'s "Framework for Responsible and Ethical Enablement of Artificial Intelligence in the Financial Sector", released on 13 August 2025 [L2-C1]. The Committee was constituted under the chairmanship of a Deputy Governor. The remit was to produce a framework for the financial sector, but the output overflowed sector boundaries. The 7 Sutras [L2-C2], the 6 Pillars and the 26 Recommendations [L2-C3] are written in general enough terms that MeitY was able to adopt them wholesale for the national AI governance regime three months later.
The second document is the MeitY "India AI Governance Guidelines: Enabling Safe and Trusted AI Innovation" released on 5 November 2025 under PIB PRID 2186639. MeitY added the 6 Pillars framing [L2-C4] across three domains (Enablement, Regulation, Oversight), articulated the three institutions (AIGG, TPEC, AISI) and set the voluntary posture. But the operational language of the Sutras is RBI language.
The practical consequence is important. If you work in the financial sector, FREE-AI is directly applicable with its 26 Recommendations as a running checklist. If you work outside the financial sector, FREE-AI is still your best interpretive key for what the MeitY Sutras mean in practice. Read FREE-AI even if you are a healthcare AI deployer or an e-commerce AI deployer, because the Sutras it defines are the ones MeitY expects you to live by.
What the 7 Sutras actually say
Sutra 1, Trust is the Foundation. Trust in AI systems must be built through demonstrable safeguards, documented transparency and accountability. "Trust" is not aspirational language in the Sutras. It is an operational test: can a reasonable regulator, a reasonable Board and a reasonable user believe the AI system is doing what you say it is doing? If no, you have not built trust yet.
Sutra 2, People First. Human-centric design, human oversight and human empowerment especially for high-impact decisions. This is the sutra that drives the human-in-the-loop requirement.
Sutra 3, Innovation over Restraint. Enable innovation; use targeted interventions where risk crystallises. This is why the Indian regime is voluntary at the national level and interventionist at the sectoral level. SEBI, RBI and IRDAI intervene where their mandate reaches; MeitY sets the frame.
Sutra 4, Fairness and Equity. AI must not entrench or amplify unfair bias. Measurement is expected. "Fair in the abstract" is not an acceptable defence; a bias-testing report is.
Sutra 5, Accountability. Accountability follows function. The developer answers for development choices. The deployer answers for deployment choices. The data provider answers for data quality. This is the single most important Sutra for an AI Governance Officer because it structures the entire programme.
Sutra 6, Understandable by Design [L2-C5]. Disclosures and explanations the intended user and the regulator can actually understand. "Technical documentation that only PhD data scientists read" fails Sutra 6.
Sutra 7, Safety, Resilience and Sustainability. AI systems must be safe for the user, resilient to attack and sustainable in operation. The sustainability part is new in the 2024-25 wave of international standards work and tracks the OECD update of 3 May 2024.
The 6 Pillars and what they operationalise
MeitY groups the 6 Pillars under three domains. Enablement covers Pillar 1 Infrastructure and Pillar 2 Capacity Building. These are the IndiaAI Mission [L2-C6] deliverables: compute, datasets, skills. For the AI Governance Officer in a deployer organisation, this domain is largely background music; you consume the public infrastructure.
Regulation covers Pillar 3 Policy and Regulation and Pillar 4 Risk Mitigation. This is where you spend most of your time. Pillar 4 Risk Mitigation structures the AI DPIA, the pre-deployment testing and the post-deployment monitoring.
Oversight covers Pillar 5 Accountability and Pillar 6 Institutions. Pillar 5 is the accountability-follows-function model operationalised. Pillar 6 is where AIGG, TPEC and AISI live. AIGG was formally constituted on 16 April 2026 and TPEC on 18 April 2026; AISI is incubated under the IndiaAI Mission with Director recruitment closed 2 June 2026 and appointment pending public confirmation as of 9 October 2026. The course flags implementation output from each as a "watch" item on your calendar.
Five failure modes in reading the stack
Reading MeitY in isolation. The Guidelines quote the Sutras in headline form; FREE-AI defines them in working form. Read both.
Treating FREE-AI as "only for banks". Non-financial deployers get the clearest operational reading of the Sutras from FREE-AI. Read it anyway.
Treating the 26 FREE-AI Recommendations as advisory wallpaper. They are the single best running checklist available in Indian AI governance literature today. Use them.
Assuming the Guidelines will become mandatory on a known date. The Guidelines are explicitly voluntary. MeitY has signalled that voluntary measures "may eventually lead to mandatory ones" without a date. Operate on a prudential schedule aligned to ISO/IEC 42001 [L2-C7] annual surveillance rhythm, not a statutory deadline.
Missing the Standing Committee on IT Report of 30 March 2026 which recommends a comprehensive AI law. If the Winter Session 2026 tables a Bill, the voluntary posture will shift. Track Parliament.
Your artifact from Lesson 2
Build a one-page tracker spreadsheet with the three columns: Instrument, Date, Our organisation\'s posture. Pre-populate with FREE-AI (13 August 2025), MeitY Guidelines (5 November 2025), DPDP Act 2023, DPDP Rules 2025, SEBI AI Advisory (5 May 2026), RBI Draft Model Risk (2026 cycle), IRDAI AI Working Group (19 June 2026), IT Rules 2026 amendment (10 February 2026), Standing Committee on IT Report (30 March 2026). Save it as Artifact 2 in your capstone workbook. Review monthly.