ITU - Global AI Governance

Overview

The International Telecommunication Union (ITU) is the United Nations specialized agency for information and communication technologies (ICTs). It coordinates the shared global use of the radio spectrum, promotes international cooperation in assigning satellite orbits, works to improve telecommunication infrastructure in the developing world, and establishes worldwide standards that ensure the seamless interconnection of a vast range of communications systems. In the realm of Artificial Intelligence (AI), the ITU serves as a neutral platform for governments, industry, and academia to build a common understanding of emerging AI technologies and address the consequent needs for technical standardization and policy guidance. The 'AI for Good' platform is the leading UN platform for inclusive dialogue on AI, co-convened with the Government of Switzerland and in partnership with over 50 UN sister agencies.

AI Governance Mandate

ITU's mandate in AI governance is rooted in its role as the UN agency for digital technologies, supported by resolutions from its Plenipotentiary Conference (e.g., Resolution 214, Bucharest, 2022) and the World Telecommunication Standardization Assembly (WTSA-24 Resolution COM4/AI). These resolutions emphasize the necessity of developing safe, secure, and trustworthy AI systems that contribute to sustainable development. Furthermore, the ITU's efforts align with broader UN initiatives, amplified by UN General Assembly resolutions like A/78/L.49 (New York, 2024), which underscore the importance of international alignment in AI governance. The ITU aims to embed principles of trust, accountability, and inclusivity into the technical standards it develops for AI.

Key AI Instruments

The ITU develops various instruments to guide AI development and deployment:

• ITU-T Recommendations (Standards): These are technical standards developed by ITU-T Study Groups. Notable examples include:
• ITU-T Y.3172: "Architectural framework for machine learning in future networks including IMT-2020" (June 2019).
• ITU-T Y.3173: "Framework for evaluating intelligence levels of future networks including IMT-2020" (February 2020).
• ITU-T Y.3174: "Framework for data handling to enable machine learning in future networks including IMT-2020" (February 2020).
• ITU-T Y.3176: "Machine learning marketplace integration in future networks including IMT-2020" (September 2020).
• ITU-T Y.3178: "Functional framework of artificial intelligence (AI)-based network service provisioning in future networks, including IMT‑2020."
• ITU-T Y.3115: Specifies architectural requirements and framework of AI-enabled cross-domain networks for future networks including IMT-2020 (February 2022).
• ITU-T X.2200-X.2249 series: Dedicated to Artificial Intelligence (AI) / Machine Learning (ML) security.
• ITU-T XSTR-SEC-AI: Technical Report providing guidelines for security management of using AI technology (May 2022).
• Focus Groups (FGs): Temporary groups that accelerate studies in strategically relevant fields, preparing the basis for standardization work. Key FGs related to AI include:
• FG-ML5G: Focus Group on Machine Learning for Future Networks including 5G (active Jan 2018 - July 2020).
• FG-AI4AD: Focus Group on AI for Autonomous and Assisted Driving (established 2020, concluded Sept 2022), focusing on behavioral evaluation of AI drivers and safety data protocols.
• FG-AI4H: Focus Group on Artificial Intelligence for Health (ITU, WHO, WIPO collaboration), working on performance benchmarking of AI for health solutions.
• FG-AINN: Focus Group on AI-Native Networks (established July 2024), exploring fundamental changes in network architecture to harness AI's potential.
• Reports and Publications: Outputs from the AI for Good Global Summits, AI Governance Dialogues, and the UN Activities on AI Report.

Implementation Mechanisms

ITU implements its AI governance objectives through several mechanisms:

• ITU-T Study Groups: These permanent groups develop ITU-T Recommendations. Study Group 13 (Future Networks) and Study Group 16 (Multimedia) are particularly active in AI standardization, with Study Group 17 focusing on security aspects.
• AI for Good Global Summit: An annual event that brings together stakeholders to discuss AI's potential for sustainable development, identify practical AI solutions, and foster collaboration on AI governance, skills, and standards.
• Regional Impact Events: Such as 'AI for Good Impact Africa', to extend the global dialogue and solutions to specific regional contexts.
• Innovation Factory and Machine Learning Hackathons: Platforms to identify and accelerate AI startups and solutions for global impact, particularly aligned with the SDGs.
• AI Skills Coalition: Supports capacity and skill development for policymakers, regulators, industry leaders, and the next generation of tech innovators, with an aim to train 10,000 people globally in 2025.

Member Engagement

ITU's approach to AI governance is highly collaborative and multi-stakeholder. It actively engages governments, industry, academia, civil society, and other UN agencies. The 'AI for Good' platform itself has a thriving community of over 37,000 active contributors from more than 180 countries. Focus Groups are open to all interested parties, ensuring broad input into the standardization process. The 'AI for Good Neural Network' is an AI-powered community and content platform designed to connect thought leaders and professionals globally, providing inclusive access to AI for Good content since 2017.

Current Focus Areas

The ITU's current focus areas in AI governance include:

• Developing standards for AI integration in 5G and future telecommunication networks, including architectural frameworks, data handling, and intelligence evaluation.
• Establishing international standards for the behavioral evaluation and safety protocols of AI in autonomous and assisted driving.
• Advancing AI for health solutions through performance benchmarking and ethical considerations.
• Addressing cybersecurity aspects of AI, including security evaluation frameworks and threat management throughout the AI lifecycle.
• Promoting environmental efficiency for AI and other emerging technologies, including model-level environmental accounting and data center efficiency.
• Working on multimedia authenticity standards, including watermarking to counter AI-generated misinformation and disinformation.
• Exploring the concept and architecture of AI-native networks where AI is deeply embedded in the core network design.

Future Directions

The ITU continues to drive the global AI conversation from principles to implementation. Future directions include sustained efforts in developing comprehensive and inclusive AI standards, expanding capacity-building initiatives, especially in developing countries, and fostering partnerships for the responsible and ethical deployment of AI. The ITU aims to ensure that AI development is aligned with the UN Sustainable Development Goals, addressing emerging challenges posed by advanced AI technologies like generative AI, and promoting global interoperability and trust in AI systems. The next AI for Good Global Summit is scheduled for July 7-10, 2026, in Geneva.

Resources