Responsible AI Framework Tracker

As AI technology matures, it’s likely that we will see the number of frameworks increase. This page consists of a collection of frameworks and international standards specifically about AI or in areas that are connected to it. The aim is to make it easier for you to choose frameworks that fit your particular use case.

Want to learn more about AI regulation that is focused on particular jurisdiction? Take a look at our free Global AI Regulations Map.

One question that you might have is how does this ‘fit’ into the work Fairly AI does? These frameworks are the kinds of documents that provide the foundation from which we build policies. Each policy is built up of individual controls which act as individual ‘questions’ about a particular issue in AI or other areas connected to it. When you use our system, you can choose an answer and provide evidence to support your answer. We then categorize our controls based on the AI lifecycle.

What this means is that for a given area, such as ‘development’ or ‘validation’, we can see where gaps exist across a number of different policies.What that tells us is that the source documents for policy (e.g. regulations, standards, and frameworks) have gaps. Since we are able to detect those gaps, we are better equipped to help you choose a series of policies that fit your use case.

We update the list below as we come across new frameworks and standards in the AI space, you can make suggestions on our GitHub and please don't forget to give us a star if you find our work useful!

General-use Responsible AI and Risk Management Frameworks

• NIST AI Risk Management Framework
• Microsoft (~221,000 employees) Responsible AI Framework
• Google (Alphabet, ~190,234 employees) Responsible AI Practices
• Salesforce (70,000+ employees) AI Ethics Maturity Model
• Rolls Royce (50,000+ employees) Aletheia Framework 2.0
• AI Ethics Impact Group led by VDE and BertelsmannStiftung (85,000 employees) - From Principles to Practice An interdisciplinary framework to operationalise AI ethics
• ISO/IEC 23894:2023 Information technology - Artificial intelligence - Guidance on risk management
• ECP Platform for the Information Society - Artificial Intelligence Impact Assessment
• Digital Catapult (Innovate UK) AI Ethics Framework
• Infocomm (Singapore) and Singapore Personal Data Protection Commission Model Artificial Intelligence Governance Framework 2e
• Office of the Privacy Commissioner of Personal Data, Hong Kong Guidance on the Ethical Development and Use of Artificial Intelligence
• Australia's AI Ethics Principles

Concept-based Frameworks

Accountability

• Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditing
• Reviewable Automated Decision-Making: A Framework for Accountable Algorithmic Systems
• UK Information Commissioner’s Office Guidance on the AI Auditing Framework
• US Government Accountability Office: An Accountability Framework for Federal Agencies and Other Entities

Benchmarking and Performance

• ISO/IEC AWI TS 29119-11 - Software and systems engineering - Software testing - Part 11: Testing of AI systems‍
• ISO/IEC NP TS 12831 - Information technology - Artificial intelligence - Testing for AI Systems (proposal)
• IEEE 2937-2022 - IEEE Standard for Performance Benchmarking for Artificial Intelligence Server Systems
• IEEE 2801-2022 - IEEE Recommended Practice for the Quality Management of Datasets for Medical Artificial Intelligence
• ‍

Agentic Systems

• Test and Evaluation Framework for Multi-Agent Systems of Autonomous Intelligent Agents
• Microsoft (~221,000 employees) 10 Guidelines for Developers of Conversational AI

Content Provenance

• The Coalition for Content Provenance and Authenticity (C2PA) Specifications

‍Cybersecurity and Safety

• MITRE ATLAS™ (Adversarial Threat Landscape for Artificial-Intelligence Systems)
• World Economic Forum Presidio AI Framework: Towards Safe Generative AI Models in Collaboration with IBM Consulting (160,000+ employees)
• NASA (17,330 employees) Hazard Contribution Modes of Machine Learning Components‍
• IEEE P7009 - Standard for Fail-Safe Design of Autonomous and Semi-Autonomous Systems
• ETSI GR SAI 006 v1.1.1 - Securing Artificial Intelligence (SAI): The role of hardware in security of AI‍
• IEEE P3157 - Recommended Practice for Vulnerability Test for Machine Learning Models for Computer Vision Applications (pre-draft)
• ISO/IEC TS 27022:2021 - Information technology - Guidance on information security management system processes
• NIST AI 100-2e2023 - Adversarial Machine Learning - A Taxonomy and Terminology of Attacks and Mitigations

Data

• ISO/IEC FDIS 5259-4 - Artificial intelligence - Data quality for analytics and machine learning (ML)
• Bias issues: Data Statements for Natural Language Processing: Toward Mitigating System Bias and Enabling Better Science
• IEEE - Data Readiness Report
• Towards Accountability for Machine Learning Datasets: Practices from Software Engineering and Infrastructure
• Data&Trust Alliance (4.6 million employed across alliance member organizations) Data Provenance Standard
• Healthcare Data Governance, Privacy, and Security - A Conceptual Framework (Faridoon and Kechadi)
• ISO/IEC 25012:2008 - Software engineering - Software product Quality Requirements and Evaluation (SQuaRE) - Data quality model
• ISO/IEC 25024:2015 - Systems and software engineering - Systems and software Quality Requirements and Evaluation (SQuaRE) - Measurement of data quality
• ‍

Definitions, Terminology, and Classification

• OECD Framework for the Classification of AI Systems
• IEEE P3123 - Standard for Artificial Intelligence and Machine Learning (AI/ML) Terminology and Data Formats
• IEEE P2802 - Standard for the Performance and Safety Evaluation of Artificial Intelligence Based Medical Device: Terminology
• ETSI GR SAI 001 v 1.1.1 - Securing Artificial Intelligence (SAI) – AI Threat Ontology
• ISO 8000-2:2022 - Data quality Part 2: - Vocabulary
• ANSI/CTA 2089 - Definitions/Characteristics Of Artificial Intelligence In Health Care

Design, UI, and UX

• Design Principles for Generative AI Applications (Weisz et al. 2024)
• CAN/CIOSC 101:2019 (R2021) - Ethical design and use of automated decision systems
• ITU-T H.862.5
• Emotion enabled multimodal user interface based on artificial neural networks
• ITU-T H.862.5 - Emotion enabled multimodal user interface based on artificial neural networks

Environmental Impact

• NWIP TR - Green and sustainable AI (N 256)
• ‍

Explainability

• UK Information Commissioner’s Office & The Alan Turning Institute Explaining Decisions Made with AI
• Advanced Technology Academic Research Center (ATARC) Artificial Intelligence — Machine Learning (ML) model transparency
• IEEE P2976 - Standard for XAI – eXplainable Artificial Intelligence – for Achieving Clarity and Interoperability of AI Systems Design (pre-draft)
• IEEE P2894 - Guide for an Architectural Framework for Explainable Artificial Intelligence

Fairness and Bias

• ISO/IEC TR 24027:2021 Information technology - Artificial intelligence (AI) - Bias in AI systems and AI aided decision making
• NIST - Towards a Standard for Identifying and Managing Bias in Artificial Intelligence
• BSA (The Software Alliance, formerly known as ‘The Business Software Alliance’) Confronting Bias: BSA’s Framework to Build Trust in AI
• National Fair Housing Alliance - Purpose, Process, and Monitoring: A New Framework for Auditing Algorithmic Bias in Housing and Lending
• ISO/IEC DTS 12791 - Information technology - Artificial intelligence - Treatment of unwanted bias in classification and regression machine learning tasks (under development)

High Impact Risk

• OpenAI (~770 employees) Preparedness Framework (Beta)

Human-Computer Interaction

• IEEE P7008 - Standard for Ethically Driven Nudging for Robotic, Intelligent and Autonomous Systems (pre-draft)
• IEEE P7014 - Ethical considerations in Emulated Empathy in Autonomous and Intelligent Systems (pre-draft)
• IEEE P3128 - Recommended Practice for The Evaluation of Artificial Intelligence (AI) Dialogue System Capabilities (pre-draft)
• IEEE 7010-2020 - IEEE Recommended Practice for Assessing the Impact of Autonomous and Intelligent Systems on Human Well-Being

Impact Assessments

• ISO/IEC DIS 42005 Information technology - Artificial intelligence - AI system impact assessment (under development)

Licensing

• IEEE P2840 - Standard For Responsible AI Licensing

Robustness

• IEEE P3129 - Standard for Robustness Testing and Evaluation of Artificial Intelligence (AI)-based Image Recognition Service (draft)
• ISO/IEC TR 24029-1:2021 - Artificial Intelligence (AI) Assessment of the robustness of neural networks

System Management

• ISO/IEC 42001:2023 Information technology - Artificial intelligence - Management system
• Designing Trustworthy AI: A Human-Machine Teaming Framework to Guide Development‍
• ISO/IEC AWI TS 5471 - Systems and software engineering - Systems and software Quality Requirements and Evaluation (SQuaRE) - Guidance for quality evaluation of artificial intelligence (AI) systems‍
• ISO/IEC 5338:2023 - Information technology - Artificial intelligence - AI system life cycle processes

Transparency

• ISO/IEC DIS 12792 - Information technology - Artificial intelligence -Transparency taxonomy of AI systems
• IEEE 7001-2021 - IEEE Standard for Transparency of Autonomous Systems

‍Trustworthiness

• European Commission - Ethics Guidelines for Trustworthy AI
• Trustworthy Artificial Intelligence Implementation (TAII) Framework
• ANSI/CTA-2090 - The Use of Artificial Intelligence in Health Care: Trustworthiness
• ISO/IEC TR 24028:2020 -Information technology - Artificial intelligence - Overview of trustworthiness in artificial intelligence
• ANSI/CTA 2096 - Guidelines for Developing Trustworthy Artificial Intelligence Systems
• VDE-AR-E 2842-61-6 - Development and trustworthiness of autonomous/cognitive systems – Part 61-6: After release of the solution
• ‍

Validation

• BS 30440:2023 - Validation framework for the use of artificial intelligence (AI) within healthcare
• ISO/IEC AWI TS 17847 - Information technology - Artificial intelligence - Verification and validation analysis of AI systems (under development)

Industry-based Frameworks

Cognitive Technology

• Cognilytica Cognitive Project Management for AI Methodology

Education

• UNESCO Guidance for Generative AI in Education and Research
• IEEE P2247.4 - Recommended Practice for Ethically Aligned Design of Artificial Intelligence (AI) in Adaptive Instructional Systems (pre-draft)
• ‍

Energy

• US Department of Energy AI Risk Management Playbook

Healthcare and Pharmaceuticals

• Roche (~103,613 employees)  Artificial Intelligence (AI) Ethics Principles
• World Health Organization - Ethics and governance of artificial intelligence for health - Guidance on large multi-modal models
• IEC SRD 63416:2023 ED1 - Ethical considerations of Artificial Intelligence (AI) when applied in the Active Assisted Living (AAL) context
• ‍

Intelligence

• Artificial Intelligence Ethics Framework for the Intelligence Community
• Principles of Artificial Intelligence Ethics for the Intelligence Community

Role-based Frameworks

Investors

• Radical Ventures' ($350M raised) Responsible AI for Startups (RAIS) framework

Boards‍

• Anekanta Responsible AI Governance Framework for Boards

Startup Founders

• Responsible Innovation Labs Responsible AI Policy (this policy applies to investors too)

Leadership and Executives

• Adobe (29,000+ employees) Empowering AI Leadership - An Oversight Toolkit for Boards of Directors‍
• IEEE P2863 - Recommended Practice for Organizational Governance of Artificial Intelligence (pre-draft)
• XP Z77-101 - Guide of good practices in matters of governance of ethical approaches within organizations

Information Technology

• ISO/IEC TS 38501:2015 - Information technology - Governance of IT - Implementation guide
• ‍

Procurement

• IEEE 3119 - Standard for the Procurement of Artificial Intelligence and Automated Decision Systems (pre-draft)
• AFNOR SPEC Z77-100-0 - Contractualization of AI systems
• ‍

Looking to operationalize a responsible AI framework? Work with Fairly to build your policy control center!

‍

Notes:

• CSET's Matrix for Selecting Responsible AI Frameworks and the AI Standards Hub provided many sources for this tracker.

‍

Changelog

• April 2024: The Coalition for Content Provenance and Authenticity (C2PA), ISO/IEC DIS 42005, IEEE P3157, IEEE P7008, IEEE P7009, IEEE P7014, IEEE P2840, ETSI GR SAI 006 v1.1.1, IEEE P3129, IEEE P3123, IEEE P3128, IEEE P2976, IEEE P2863, IEEE P2802, IEEE P2247.4, ISO/IEC TR 24029-1:2021, ISO/IEC 5338:2023, ISO/IEC TS 25058:2024, ETSI GR SAI 001 v 1.1.1, ISO/IEC 25012:2008, ISO/IEC 25024:2015, ISO 8000-2:2022, ANSI/CTA 2089, BS 30440:2023, ISO/IEC DTS 12791, ISO/IEC DIS 12792, ISO/IEC AWI TS 29119-11, ISO/IEC TR 24028:2020, ISO/IEC NP TS 12831, IEEE 7001-2021, ANSI/CTA 2096, XP Z77-101, NWIP TR N256, CAN/CIOSC 101:2019 (R2021), IEC SRD 63416:2023 ED1, ITU-T H.862.5 (06/2021), ISO/IEC TS 38501:2015, ISO/IEC TS 27022:2021,

• March 2024: Anekanta Responsible AI Governance Framework for Boards, Design Principles for Generative AI Applications,
• ‍February 2024: Australia's Responsible AI Framework, OpenAI's Preparedness Framework (Beta), UNESCO's Guidance for Generative AI in Education and Research, Radical Ventures' Responsible AI for Startups (RAIS) framework, Responsible Innovation Labs Responsible AI Policy.

‍

‍

DISCLAIMER. The insights and information contained in this article is intended for informational purposes only and we at Fairly AI do not intend for it to constitute legal advice, opinion, or recommendations that you, the reader, or others can depend on. For any legal issues linked to the information and insights contained in this article, please contact a lawyer/attorney to retain legal advice.

Fairly provides links to other websites beyond our control and we are not responsible for and do not give any warranties or make any representations regarding such websites. We are not responsible for or liable in relation to the content or security of these sites (other than to the extent required by law.)

Fairly makes reasonable endeavors to ensure reliability of the information presented on the app and website, but makes no warranties or representations whatsoever as to its accuracy or reliability. Similarly, due to the nature of the internet, we do not guarantee that the Fairly will be available on an uninterrupted and error free basis.

‍