ITU is the leading international organisation to ensure sustainable digital transformation and universally connecting the unconnected. The Telecommunications Standardization Sector ITU-T plays the key role in the development of international standards in the area of telecommunications and ICT. Today, artificial intelligence has the most serious influence on the entire telecom industry: on one hand, it creates new opportunities – such as dynamic management of spectrum and multi-satellite constellations; on the other hand, AI is becoming a spawner of new threats, first of all in the area of information safety and data security. In this connection, the ITU cannot leave the issues of regulation and application of AI unattended. Acting as UN’s specialised institution in the field of digital technologies ¹, ITU plays the key role in the global regulation and development of artificial intelligence (AI). ITU’s activities in this sphere cover the development of standards and regulatory framework, as well as creation of conditions for the practical application of AI for the achievement of the UN Sustainable Development Goals (SDGs).

Principles and Regulatory Documents

The activities of ITU in the AI regulation and development are holistic and systematic in nature. The Union acts as the key global platform where technical standards are elaborated, regulatory dialog is maintained, ethic principles are promoted, and practical initiatives are implemented. The main objective is to ensure that the development of AI is safe, inclusive, and benefits the whole humankind, assisting in the achievement of the UN’s SDGs.

Today, in its AI research, ITU is guided by the following documents:

• Resolution 214 ² (Bucharest, 2022), which defines general directions of the ITU’s work on AI, particularly where AI impacts the development of the telecommunications and ICT sector. The primary objective is to create a reliable telecom/ICT ecosystem. The ITU should facilitate the exchange of information and expertise, coordinate efforts, and promote cooperation;

• Resolution 101 ³ (New Delhi, 2024), which goes specific on the activities of the ITU Standardization Sector (ITU-T) in streamlining AI technologies and for the support of telecommunications and ICT.

The ITU keeps collecting information among the member states on their initiatives in the field of policy and AI regulation and on how that work is correlated with their efforts in digital transformation. These matters have become part of the ITU’s research on regulatory issues ⁴.

AI Standardisation

One of the key lines of work for the ITU to ensure safe and reliable development of AI is standardisation.

Arnaud Taddei, chairperson of ITU-T Study Group 17, the organisation’s expert taskforce for the elaboration of standards in the area of safety, cybersecurity and trust management, compares ⁵ the challenge of standardising the agentic AI with the development of the Open Systems Interconnection (OSI) model as basis for hardware and software systems which support global communication networks today. Group 21 (SG21) has developed Recommendation ITU-T F.748.56 "Artificial intelligence generated content – General framework and requirements)⁶. Study group No. 5 (SG05) is working on the environmental assessment of AI.

AI Solving Problems of Humanity

The ITU is actively promoting AI to solve global problems. The AI for Good platform⁷ is the ITU’s flagship initiative launched in 2017. This is a multilateral community uniting governments, private businesses, academia and social organisations (including UN institutions) for the responsible development of AI for the benefit of sustainable development. The major goal of the project is to help the world harness the potential of artificial intelligence to solve global problems and to accelerate the progress in achieving the UN’s Sustainable Development Goals. The project community comprises over 55,000 members from 180 countries.

The platform is host to the AI for Good Global Summit and events like AI Governance Day for discussing development strategies. The annual Global Summit in Geneva is the premier event connecting AI innovators with those seeking solutions to global challenges. This year, it will take place from 6 to 10 July .

Development of AI standards and policies to boost innovations and digital transformation. Educational initiatives (such as AI Skills Coalition) aimed at improving AI literacy and training of future specialists. Hands-on programs such as Innovation Factory startup accelerator, machine learning competitions, pilot zones for the testing of standard solutions.

The forum is joined by the leading players of the information and communication industry – public, commercial and academic structures: Amazon, Meta, Google, Microsoft, Salesforce, NVIDIA, Warner Bros, Deloitte, World Bank, UNESCO, SpaceX, Universal Music Group, UNOOSA, and others. One of the main speakers is Ms Doreen Bogdan-Martin, Secretary-General of the ITU.

Intersputnik is a member to the ITU’s two key sectors: Radiocommunication Sector (ITU-R) and Telecommunication Development Sector (ITU-D), which allows the Organization to take an active part in all constructive initiatives of the ITU. In November 2025, Intersputnik and the ITU signed an agreement on the development of a platform for training and professional development of employees in the area of satellite communications and broadcasting. Both the satellite communication development and the training of employees are impossible without AI today, so Intersputnik not only tracks development trends in this industry but also actively participates in shaping them.

Global Symposium for Regulators

Also, the ITU conducts the Global Symposium for Regulators (GSR): An annual event where national regulators discuss new regulatory horizons. Since 2018, the GSR has featured special sessions dedicated to AI, covering ethical aspects, standardisation, and risk mitigation.

The ITU strives for inclusivity, so that all countries can use the benefits of AI. The ITU emphasises the importance that all voices, especially those from developing countries, are heard in the global AI-concerned dialog.

GenAI for Good Challenge

GenAI for Good Challenge is a global contest organised by the International Telecommunication Union (ITU) jointly with IEEE Humanitarian Technologies ⁸. Its aim is to find and support practical solutions based on generative artificial intelligence (GenAI) to solve the real problems in healthcare, agriculture, and climatic sustainability; searching for applied solutions for specific tasks in the area of healthcare, agriculture, and climate on the local level. All projects should be aimed at achieving the UN’s Sustainable Development Goals.

The contest opened in 2025 focuses on three key areas, each of them including a specific pilot project in a certain country:

• agriculture (Lesotho): development of a mobile chatbot with voice interface to provide farmers with advice, forecasts, and market updates;

• healthcare (Gambia): creation of a multi-language chatbot for the prevention of non-infectious diseases;

• climate (Bangladesh): development of an AI assistant for farmers with localised forecasts and advice on getting ready for draughts.

More than 300 applications arrived from 79 countries. Right now, the winners of the preliminary stage are deploying their pilot projects in real life conditions. The results will be summarised in July 2026.

AI and Radio Broadcast

Radio broadcasting remains important thanks to the trust of the audience to certain hosts and also owing to the infrastructure that is resilient to external disruptions. The future of the radio is dependent on the adequate use of the AI, which helps create new content and automatise processes: transcription, compilation of playlists, multi-language translation, and advertisement. At the same time, it must be recognised that AI will not replace DJs and presenters; human involvement in radio broadcasts remains an invariably crucial factor⁹.

Therefore, it is necessary to adequately manage AI so as to protect the audience and to maintain their trust.

AI may improve the quality of early warning. Mobile network is the first to face problems in case of emergencies. Radio network is resilient to destruction of infrastructure and remains an accessible source of information, covering a large audience. Radio remains important for notifying people after the restoration of the mobile network. The sixth study group of the ITU (SG 6) handles the broadcasting issues, including the use of AI in the broadcasting: the effect of AI on production, assessment of quality and delivery of the content. The objective of the group is to ensure functional compatibility of systems all over the world.

Managing Radio Frequency Spectrum With AI

Artificial intelligence is increasingly making inroads into RF spectrum management, offering solutions to improve efficiency and automation. The ITU as global regulator in this area plays the key role in studying, standardising and implementing these technologies.

AI and machine learning help to:

• predict spectrum availability: analysing massive datasets to accurately assess and forecast which frequencies are vacant at present and in the near future;

• automatise the monitoring: process data in real time, detect interferences and anomalies without continuous human intervention;

• dynamically manage the spectrum: allocate frequencies in real time and adapt the allocation to the current network conditions, which allows for supporting more users without losing quality;

• create intelligent radio systems: such systems are able to autonomously identify free spectrum and switch over to other frequencies, and also to detect and classify interferences on their own.

The activities of the ITU in this field cover several key directions:

• development of standards (Radiocommunication Sector, ITU-R): study group SG 1 is working on the “ITU-R SM.[AI AND BIG DATA] - Spectrum Monitoring - Proactive, Autonomous, and Data-Driven” report¹⁰, which is dedicated to the next generation of AI- and big data-driven spectrum monitoring systems;

• practical application and pilot projects: The ITU supports the development of practical tools. One of the examples is the SIMONES¹¹ system, which is positioned as a low-cost spectrum monitoring system for developing countries. The ITU is placing great emphasis on AI as a factor in eliminating the digital divide. Implementation of AI-based solutions for the spectrum management will help developing countries use the limited frequency resources with maximum efficiency;

• political and regulatory support: The ITU assists in integrating AI principles with global regulatory processes, first of all through the work at World Radiocommunication Conferences (WRCs). Decisions made at the WRCs outline the way the spectrum is used by technologies that increasingly support AI applications. Today, the ITU is exploring the possibility of using large language models (LLMs) to speed up and automatise bureaucratic procedures related to spectrum regulation.

ITU and AI regulation

At GSR-26 (Global Symposium for Regulators) (held from 12 to 15 May, Ankara, Turkey), AI regulation became one of the central topics. The main result was the adoption of a new international guide, "2026 Best Practice Guidelines: Regulatory Governance Essentials»¹² – a set of tools for regulation in the quickly changing digital environment This document defines a new vector for digital regulation, moving away from rigid and uniform norms in favour of more dynamic approaches.

It is proposed that the strictness of regulation depend on the size, technological capabilities, and level of risk posed by a particular digital entity (operator, platform, AI service). This makes it possible to avoid applying the same requirements to all market players.

Instead of rigid and quickly outdated rules, it is recommended to implement adaptive models and use regulatory “sandboxes” to implement pilot projects in. This approach allows norms to keep pace with developments in generative AI and other technologies.

The document also opens the door for regulators to use AI tools themselves to analyse data and monitor markets. This makes surveillance more responsive and data-driven.

In addition to the adopted document, various aspects of AI regulation were actively discussed at the symposium.

The regulators have came to a conclusion that their role is shifting from simply responding to technological change to proactively designing regulatory pathways that foster sustainable and inclusive ecosystems. The main goal is to set the conditions for innovation and investment.

The sessions explored specific regulatory pathways for emerging areas such as AI governance and digital innovation.

The dual nature of the technology was discussed. On the one hand, the enormous economic potential (estimated at $4.8 trillion in contribution to the global economy) and the speed of its implementation were noted. On the other hand, questions were raised about risks such as deep fake fraud and algorithmic bias.

Particular attention was paid to the role of AI in ensuring network resilience, for example, to maintain communications during emergencies and improve the effectiveness of early warning systems.

The importance of high-quality data for AI-based analytical platforms, which help regulators better monitor markets, was discussed. The need to consider the accuracy of results when using AI for predictive analytics in regulatory processes was also emphasised.

The special session “Network of Women (NoW) in ITU-D” discussed inclusive approaches to regulation, including in the emerging area of ​​AI governance, as well as strategies to extend women’s influence in regulatory decision-making.

ITU Secretary-General Doreen Bogdan-Martin stressed¹³ that despite the rapid growth of AI, almost a third of the world’s population (2.2 billion people) still do not have access to the Internet. In this regard, the task was set to prevent the existing digital divide from turning into an even more severe AI divide.

This statement is not a random phrase – it reflects the systemic position of the ITU. Ms Bogdan-Martin has highlighted this issue several times already: at the India AI Impact Summit¹⁴, at the G20 meeting in South Africa¹⁵, and at the UN Human Rights Council¹⁶.

Ms Secretary-General consistently points to three key aspects of the problem:

• scale of the lag: 2.6 billion people still lack internet access, and 16% of countries have significant computing power, while entire continents have almost none;

• deepening inequality: AI technologies are advancing at an unprecedented pace, potentially leaving lagging countries and communities far behind;

• need for action: To prevent this, the ITU is focusing on three areas: standards (for trust and compatibility), skills (so that people can use AI), and solutions (to make a real impact on people’s lives).

Thus, the regulatory focus of GSR-26 was to find a balance: encouraging innovation and investment in AI while managing the associated risks and ensuring that the benefits of new technologies are accessible to all.

On July 6–7, the Global Dialogue on AI Governance will take place in Geneva¹⁷ – a key step in the implementation of the Global Digital Compact, the first universal agreement on digital cooperation and AI governance.

Another important ITU event in the field of AI regulation – the World Telecommunication/ICT Policy Forum (WTPF-26)¹⁸ – will take place from 2 to 4 September in Nassau and will become a platform for exchanging views and shaping up a common pathway to understanding the regulatory policy, also for artificial intelligence.

The main topics of WTPF-26 are: accelerating the creation of an inclusive, sustainable and innovative digital future, bridging the digital divide, green digital transformation, resilience of ICT infrastructure, space connectivity, strengthening ICT-focused innovation ecosystems and entrepreneurship. Bridging the digital divide requires global collaboration between governments, the private sector, academia, technology, and all of civil society.

Satellite communications will also be discussed as an important tool for the implementation of digital technologies, which is particularly relevant for small island developing states.

AI and UNESCO

In 2018, UNESCO began working on the ethics of artificial intelligence¹⁹. In 2021, UNESCO’s recommendation was endorsed by all 193 Member States. More than 80 countries have already implemented the UNESCO’s recommendation.

AI in Earth Observation industry

Today, AI is transforming traditional satellite survey into “Earth Intelligence”²⁰, a concept actively promoted by the Group on Earth Observations (GEO)²¹.

Before we can talk about the full-fledged use of monitoring results on a global scale, several issues need to be addressed.

• Unequal access to data: Thousands of satellites operate in remote sensing systems, but processed information (decision-ready products) is not available to all countries and users.

• Complexity of extracting semantic information: Raw images require complex algorithms to identify meaningful features (e.g. crop type or flood zone).

• Ground validation: AI processing results need to be verified against ground-based data, which is often hindered by national limitations and lack of infrastructure.

• Standardisation: For global applications, unification of processing standards, models and algorithms is necessary, otherwise data from different sources are not comparable.

Earth Intelligence assumes the existence of shared computing capacities instead of national processing centres in each country. And this leads to the existence of international hubs that allow users to process images for their professional needs without in-depth technical knowledge.

Another element of Earth Intelligence is Edge AI, which makes it possible to perform basic processing directly on a smartphone or other non-professional device, which is critical for remote areas without a stable high-speed internet connection. This approach significantly reduces the barrier to entry, which is especially important for developing countries.

Political and Regulatory Aspects

Earth Observation AI within the Earth Intelligence concept will require specialised regulation that goes beyond general AI governance rules. For example, it will be necessary to coordinate temporal and spatial data - coordinates and precise time scales – at the international level. Otherwise, data from different countries will be incompatible.

National restrictions on the use of strategically significant socio-economic data (e.g. crop yield statistics or post-disaster damage assessments) that are integrated with satellite imagery will need to be considered.

Today, the GEO Global Forum is already in dialogue with the ITU on the development of such specialised rules.

Onboard AI systems conserve spectrum by filtering relevant information, reducing congestion on communication channels, and by not transmitting raw data to Earth. AI processing in space is currently being used in pilot mode; its large-scale application will require solutions to radiation resistance issues.

Thus, AI is becoming a critical factor in many aspects of human activity, a mechanism for solving humanity’s most pressing problems – for example, connecting the unconnected universally. Without widespread use of AI, achieving the UN’s Sustainable Development Goals is not possible. However, AI development must proceed within an adequate regulatory and standardisation framework to ensure safe and equitable access to AI applications for all humanity. Standardisation, regulation, and the refining of inclusive solutions, primarily suitable for developing countries, are the main areas of the ITU’s work in the field of AI today.

The AI ​​for Good global forum in Geneva is expected to answer many questions about AI development, its regulation, and equitable use to achieve the UN’s SDGs. This forum will discuss current problems in AI development and identify areas for further research. The ITU remains the primary regulator of this industry, developing rules that encourage equitable international cooperation to create a modern digital ecosystem.