AI Governance

The world.s first complete AI regulation. Uses a risk-based tier system: unacceptable risk (banned), high-risk (strict requirements), limited risk (transparency duties), and minimal risk (no obligations). Phased enforcement through 2026.

Voluntary framework for managing AI risks across design, development, deployment, and use. Organized around four functions: Govern, Map, Measure, Manage. Intended as a living document with regular reviews.

Cross-sectoral profile of the AI RMF 1.0 specifically for Generative AI. Defines 12 risk categories and ~200 actions for managing GenAI risks across the AI lifecycle.

First global standard on AI ethics, adopted by 193 Member States. Ten core principles centered on human rights, dignity, and environmental protection. Covers data governance, gender, education, and health policy action areas.

First intergovernmental standard on AI. Five values-based principles and five national policy recommendations. Updated in 2024 to reflect generative AI developments. Basis for the G20 AI Principles.

Voluntary code of conduct for organizations developing advanced AI systems, including foundation models and generative AI. Builds on OECD AI Principles. Covers safety, security, transparency, and accountability throughout the AI lifecycle.

Five principles to protect civil rights in the age of AI: safe and effective systems, algorithmic discrimination protections, data privacy, notice and explanation, and human alternatives. Accompanied by a technical companion with implementation guidance.

World-first agreement signed by 28 countries and the EU at Bletchley Park. Establishes shared understanding of frontier AI opportunities and risks. Commits signatories to international cooperation on AI safety research and risk management.

Building on the Bletchley Declaration, 10 countries and the EU commit to interoperable risk-based AI governance frameworks. Supports AI safety institutes, the Hiroshima Code of Conduct, and frontier AI safety commitments.

Final report proposing seven recommendations for international AI governance. Addresses the global governance deficit, calls for a light institutional mechanism, and urges equitable access to AI benefits. Based on consultations with 2,000+ participants across all regions.

First international legally binding treaty on AI. Opened for signature in Vilnius. Ensures AI activities are consistent with human rights, democracy, and the rule of law. Covers both public and private actors. Technology-neutral design.

Practical guidance for organizations to deploy AI responsibly. Translates ethical principles into concrete recommendations covering internal governance, decision-making, and stakeholder communication. Updated with industry examples.

Expands the Model AI Governance Framework for generative AI. Nine dimensions addressed in totality: accountability, data, development, testing, deployment, incident response, security, content provenance, and user empowerment.

Binding regulation for generative AI services in China. Requires content moderation, algorithm registration, data provenance, and labeling of AI-generated content. Applies to services provided to the public within the PRC.

Comprehensive executive order directing federal agencies on AI safety, security, and trustworthiness. Mandates red-team testing for frontier models, NIST standards development, content provenance, and biotechnology risk assessment.

Standard for integrating ethical values into system design. Provides processes for ethical value elicitation, prioritization, and traceability through operational concepts, value propositions, and risk-based design. Available free via the IEEE GET Program.

World's first AI management system standard. Specifies requirements for establishing, implementing, maintaining, and continually improving an AI management system (AIMS). Applicable to any organization providing or using AI products or services.

Guidance on managing risks specific to AI. Helps organizations integrate AI risk management into their activities and functions. Describes processes for effective implementation and integration. Complements ISO 31000 and ISO/IEC 42001.

Conceptual framework and shared terminology for describing AI systems that use machine learning. Defines components and functions of ML-based AI systems. Foundation for more specific standards on risk management, governance, and conformity assessment.

ISO/IEC adoption of IEEE 7000-2021. Provides processes for including ethical values throughout system concept exploration and development. Supports traceability of ethical values through operational concepts, requirements, and risk-based design.