As the global tide of artificial intelligence (AI) innovation continues to rise, the AI for Good Global Summit 2025, organised by the International Telecommunication Union (ITU), has reaffirmed the importance of inclusive collaboration in shaping AI governance, advancing technical standards, and building capacity to ensure that AI serves the broader good of humanity.
The four-day summit, held under the auspices of the United Nations’ digital technologies agency, brought together government leaders, international organisations, civil society groups, academia, and the private sector to deliberate on building frameworks that support responsible and sustainable AI development. Discussions and showcases highlighted AI’s transformative role across various sectors, including healthcare, food systems, energy efficiency, road safety, and multimedia authentication.
This year’s event placed strong emphasis on global AI governance through an AI Governance Dialogue led by the ITU and 53 United Nations partners. High-level government officials and global stakeholders convened to identify pathways for translating AI principles into actionable policy frameworks that drive equity, transparency, and innovation.
In a co-chairs’ summary report presented during the summit, stakeholders outlined a shared vision for AI that bridges innovation with inclusion and sustainability. The report distilled key insights and reaffirmed the collective resolve to turn dialogue into coordinated action.
The co-chairs highlighted the need to move “from principles to practice”—emphasising that AI governance must go beyond lofty ideals to embrace practical implementation, backed by agile regulatory frameworks, inclusive oversight, and adaptable technical tools that can guide AI’s development in socially and environmentally responsible directions.
They underscored the necessity of multistakeholder involvement, noting that effective governance must be co-created by governments, civil society, academia, international organisations, the private sector, and technical communities. Ensuring that developing countries have a seat at the table, supported by capacity-building initiatives, is vital for equitable global AI benefits.
Transparency was identified as a cornerstone of trust in AI systems. Understanding how AI tools are built, evaluated, and used—especially in terms of model behaviour, data practices, and decision-making processes—was described as crucial to fostering accountability and public confidence.
The report also stressed the importance of inclusion, calling for AI governance to reflect diverse global perspectives. Inclusion should not stop at providing access but must extend to enabling meaningful participation in the development of the technologies and rules shaping people’s daily lives.
Closing gaps in AI preparedness through capacity-building was highlighted as a critical focus. The co-chairs called for efforts across policy development, skills training, financial support, and institutional strengthening to empower communities especially in developing regions—to innovate responsibly in vital sectors such as health, education, and agriculture.
On environmental sustainability, the co-chairs noted the need for governance frameworks that reduce AI’s environmental footprint by integrating energy and environmental policies, promoting efficient data centre usage, and scaling AI infrastructure without overwhelming local systems.
They also called for sector-focused and cross-society collaboration, recognising that AI’s impact is most effectively realised in its application to real-world challenges such as disaster management, healthcare delivery, and education. Governance efforts, they argued, must engage relevant communities and align with international development frameworks.
To ensure technical safety and interoperability, the report pointed to the importance of globally coordinated technical standards, benchmarks, and audit protocols. These tools, developed through multistakeholder processes, should be adaptable to fast-paced AI innovation while maintaining safety and reliability.
The co-chairs further highlighted the emerging importance of governing compute resources and foundation models, given the growing scale and complexity of AI systems. Access to compute infrastructure, comprehensive risk assessment protocols, and accountability mechanisms were outlined as essential for ensuring that advanced AI technologies serve the public interest.
Finally, the report stressed the importance of policy coherence and agility, advocating for governance that integrates adaptable regulations with inclusively developed technical standards. Such frameworks would enable countries to stay responsive to technological evolution while avoiding fragmentation of global AI policy.
Key outcomes from the AI for Good Global Summit 2025 include several notable initiatives and deliverables aimed at advancing AI governance and impact:The launch of the Global Initiative on AI for Food Systems, a collaboration between ITU, the UN Food and Agriculture Organization (FAO), the World Food Programme (WFP), and the International Fund for Agricultural Development (IFAD), seeks to harness AI to improve agricultural productivity, efficiency, and food security on a global scale.
Two significant resources on standards and policy for multimedia authenticity were unveiled by the AI and Multimedia Authenticity Standardisation Collaboration, led by ITU in partnership with the International Organisation for Standardisation (ISO), International Electrotechnical Commission (IEC), and other standardisation communities. These resources aim to combat the growing threat of deepfakes by advancing tools for verifying multimedia authenticity and provenance.
A newly introduced AI Standards Exchange Database was also announced. This platform, currently featuring standards from ITU, ISO, IEC, IEEE, and the Internet Engineering Task Force (IETF), is designed to facilitate harmonised development and application of AI standards. The database aims to improve coordination among standards developers and assist stakeholders in adopting comprehensive and interoperable standards suites.
Reflecting on the summit’s impact, Doreen Bogdan-Martin, secretary-general of the ITU, said, “This year’s AI for Good Global Summit has shown that when people come together, dialogue can lead to inclusive cooperation, concrete action, and hope. Let’s recommit to treating AI not as an end, but as a means to do good, for the benefit of all humanity, everywhere.”
