All instruments Policy India: FREE-AI Committee Report: Framework for Responsible and Ethical Enablement of Artificial Intelligence No summary is recorded for this instrument. Status unknown Not yet reviewed English Open the source (opens in a new tab) Suggest a correction Details Full text Details Identity Title FREE-AI Committee Report: Framework for Responsible and Ethical Enablement of Artificial Intelligence Original title Not recorded Instrument ID in-rbi-free-ai-framework-2025 Type Policy Status Status unknown Jurisdiction and publisher Jurisdiction India Publisher Reserve Bank of India (RBI) FREE-AI Committee Language English Dates and versions Date August 13, 2025 Versions Kind not recorded · August 13, 2025 Classification Policy areas Not recorded Source Source URL https://naavi.org/uploads_wp/2025/AI_framework_rbi_aug_2025_report.pdf Retrieved August 29, 2026 License Unknown. Follow the publisher’s terms. Access note Not recorded Full text 197,426 characters Original text, English · 197,426 characters in 2 parts Load the full text Download as .txt Jump to part Jump to part Part 1 of 2Part 2 of 2 Outline (122 headings) Chapter 1 – Introduction and Background 1.1 Evolution of Artificial Intelligence and Machine Learning 1.2 AI and ML in Financial Services 1.3 Constitution of the Committee 1.4 Terms of Reference 1.5 Methodology 1.6 Structure of the Report Chapter 4 presents the Committee’s proposed Framework for Responsible and Ethical Chapter 2 – AI in Finance: Opportunities and 2.1 Benefits and Opportunities 2025 World Economic Forum white paper5 on AI in Financial Services, projected 2.2 Emerging Risks and Sectoral Challenges Chapter 3 – AI Policy Landscape and Insights from 3.1 Global Policy Developments and Approaches 3.2 India’s Policy Environment and Developments 3.3 Insights from Surveys and Stakeholder Engagements Chapter 4 – Building a Responsible and Ethical AI 4.1 Trust as the Cornerstone 4.2 Enablers and Considerations for Advancing Trustworthy AI 4.3 The Seven Sutras - Guiding Principles 4.4 Principles to Practice - Recommendations 4.5 Conclusion - Weaving It All Together 3 Innovation over Restraint: Foster responsible innovation with purpose 4 Fairness and Equity: AI outcomes should be fair and non-discriminatory 5 Accountability: Accountability rests with the entities deploying AI 6 Understandable by Design: Ensure explainability for trust 6 Committee’s 7 Sutras to provide flexible, forward-looking 1 Regulation, RBI General Manager-in-Charge, and 05, 2025 2 Supervision, RBI General Manager, and team 05, 2025 3 Council of India and CEO, and team 05, 2025 5 Certified Sundeep Jakhar, Head of Public 10 Council of India Officer 2025 12 Legal Policy Law and Technology and Ms Vrinda 2025 16 India Director and Head of Digital Banking 2025 19 Shukla Science, Indian Institute of 2025 1 AU Small Finance Bank Limited 30 Kotak Mahindra Bank Limited 2 Axis Bank Limited 31 Lendingkart Finance 4 Bandhan Bank Limited 33 Navi Finserv Limited 5 Bank of Baroda 34 Neokred 6 Bank of India 35 One Mobikwik Systems Limited 7 Bank of Maharashtra 36 OpenAI Inc 9 Canara Bank 38 Perfios Software Solutions Private Limited 11 Central Bank of India 40 Punjab & Sind Bank 14 Easebuzz Private Limited 43 Sammaan Capital Ltd 15 Epifi Technologies Private Limited 44 Samunnati Finance Private Limited 16 Ernst & Young LLP 45 Shriram Finance Limited 17 Federal Bank Limited 46 Signzy Technologies Pvt Limited 18 HDB Financial Services Limited 47 Slice Small Finance Bank Limited 19 Hero Fincorp Limited 48 SMFG India Credit company Limited 20 HSBC Limited 49 South Indian Bank Limited 21 ICICI Bank Limited 50 Standard Chartered Bank 22 IDBI Bank Limited 51 State Bank of India 23 IDFC FIRST Bank Limited 52 Tata Consultancy Services Limited 24 Indian Bank 53 UCO Bank 25 Indian Overseas Bank 54 Uni Cards 26 IndusInd Bank Limited 55 Union Bank of India 27 JPMorgan Chase Bank NA 56 YES Bank Limited 28 Karnataka Bank Limited 57 Yubi 29 Karur Vysya Bank Limited 58 Zerodha Capital Private limited 1. RBI (i) Accountability for (i) Specific clauses addressing AI 2. Cyber (i) Data confidentiality, (i) May include AI specific threats 3. Guidelines (i) Data privacy and consent (i) May include providing 4. Master (i) Customer rights and (i) May include awareness to 5. Master (i) Framework for early (i) May encourage AI driven fraud 6. Master (i) Broad IT governance (i) May include Access Control 7. Master (i) Risk assessments and (i) May be required for the service 1 Purpose and • Define the role of AI in the organisation 2 Principles Set out the principles in alignment with the FREE-AI 4 AI Life cycle • Ensure rigorous testing and internal product 7 Third-Party Define the scope, responsibility, and liability of third-party 8 Policy Review The AI policy may be reviewed at least once annually to 1. General Information 2. Incident Summary 3. Additional information 1 Accountability 4 Agentic AI 7 AI inertia 8 AI inventory 9 AI Safety Institute 10 Algorithmic trading 13 Auditability 14 Behaviour Audit 15 Black box problem 16 Credential stuffing 17 Data minimisation 18 Data poisoning 19 Decision trees 20 Deepfake places, entities or events and would falsely appear to a Article 3(60)]: A subset of ML using deep neural networks to mimic 21 Deep learning 22 Differential privacy 25 Equity Fair treatment of individuals. [Merriam-Webster] 27 Explainability influencing the AI system results in a way that humans 28 Fairness 29 Federated learning across a distributed network trains a global model using 30 Fine tuning Adapting pre-trained models for specific tasks. [IBM] 31 Foundation models 32 Generative AI 34 Hallucination that produces outputs that are inaccurate and 37 Inference attacks attributes of a training data record, given partial 39 Landing zones 41 Logistic regression 42 Machine learning 43 Memory scraping memory, such as payment card details, login 44 Metadata 45 Model Bias 48 Model Correlation measures how similarly or differently the models 49 Model Degradation performance over time, resulting in less accurate or 50 Model Distillation learnings of a large pre-trained model, the “teacher 51 Model Drift 52 Model Herding due to relying on shared data or design, leading towards 53 Model Life Cycle 55 Model Inversion 56 Model Risk 57 Model Validation 59 Neural Networks 62 Prompt Injection 63 Red teaming 64 Augmented relevant information from external knowledge bases and 67 Synthetic data model that is trained to reproduce the characteristics 68 Transparency relevant stakeholders in a comprehensive, accessible 70 Understandability The outline comes from headings detected in the text, so it might miss some or include extras. JavaScript is turned off. To read the text, open each of its 2 parts: Part 1Part 2 300 of 1,262, newest first Previous DETECT Act of 2025 Next Deceptive Practices and Voter Intimidation Prevention Act of 2025 Show a random instrument All instruments