The Preliminary Report by the Independent International Scientific Panel on Artificial Intelligence provides an evidence-based assessment of the risks and opportunities presented by a defining technology of our time.

“Concentrated commercial and geopolitical interests are largely dictating the speed and direction of the development and deployment of AI,” said Panel Co-Chair Yoshua Bengio. “We need a coordinated international and democratic approach with science and compassion remaining as our compass in navigating AI.”

As Co-Chair Maria Ressa put it: “Without facts you can’t have truth. Without truth you can’t have trust.”

The panel report, based on findings by 40 leading experts from academia, industry, civil society and the technical community, laid vital groundwork for the first Global Dialogue on AI Governance called by the United Nations on 6-7 July.

UN Secretary-General António Guterres said the key takeaways were about the speed of AI development, the concentration of AI power, and the persuasiveness of AI-enabled falsehoods and deception.

“A technology that can reshape economies, transform the world of work, sway elections, and tilt the balance of security is being deployed faster than anyone – including the people building it – can keep up,” he said.

The AI Dialogue addressed the Scientific Panel’s findings through discussions in four thematic clusters.

AI opportunities and implications

The hype around new and ever-improving AI models increasingly obscures the actual value of AI systems, suggested panellist Lacina Koné, CEO of Smart Africa, an alliance of 42 countries and organizations promoting digital transformation.

“People don’t want better AI, they want better jobs, better health, better education and public services,” he said. “AI is just a tool.”

The benefits of AI cannot be realised unless local languages, cultures and needs are addressed. “As AI is evolving, it is increasingly important to understand the cultural nuances, and make sure the AI experience is serving each community, including the most vulnerable,” said Yossi Matias, Vice President of Google and Head of Google Research.

But AI’s success or failure will also, ultimately, hinge on how well it co-exists with our natural environment. That will require curbing AI-related energy use, greenhouse gas emissions, and water and material consumption.

“The future of AI cannot be separated from the future of the planet,” observed Sally Radwan, Chief Digital Officer at the United Nations Environment Programme (UNEP).

Bridging AI divides

With nearly one quarter of humanity still not using the Internet, people’s chances of gaining AI proficiency are extremely unevenly distributed.

Countries need long-term investment in education, research, institutions, standards, financing and local innovation so that they can create, adapt and govern AI, not simply consume technologies developed elsewhere.

Developing countries need help with evaluations, procurement standards, and incident reporting for AI.

“One of the least-discussed AI divides is the lack of oversight capacity,” said Urvashi Aneja Founding Director of Digital Futures Lab, an independent interdisciplinary research studio based in Goa, India.

Safe, secure and trustworthy AI

Current AI safety systems are often developed without evidence and lack sufficient human oversight, transparency and accountability. Such safeguards are needed throughout the entire system lifecycle, not just when models are trained, panellists noted.

“Effective governance is not uniformity, but the ability to connect different approaches through shared standards, evidence and cooperation, and ensuring meaningful cooperation across nations,” said Costa Rica’s Minister of Science, Innovation, Technology and Telecommunications, Paula Bogantes Zamora, who co-chaired the AI safety discussion.

Respecting, protecting and promoting human rights

Increasingly familiar AI impacts include threats to privacy, safety, due process, non-discrimination and equal treatment before the law, particularly as it gathers, exploits and manipulates people’s personal data.

Children, women and disadvantaged groups face specific risks, including unsafe AI companions, discriminatory outcomes and harms that are not yet adequately recognized or addressed by law, AI Dialogue panellists said.

Human rights principles must therefore be embedded into AI systems as concrete obligations, safeguards and red lines, with clear accountability across the AI life cycle.

Affected communities, including children and young people, must have a meaningful voice in decisions about how AI is designed, deployed and governed.

Even well-meaning AI governance discussions are disproportionately concentrated in the European UN hub, Geneva, noted Linda Bonyo, CEO of the Lawyers Hub, a Nairobi-based digital law organization.

What the UN was built for

The UN was founded 81 years ago to solve common economic, social, cultural and humanitarian problems collectively.

“The UN was built for moments like this,” said Annalena Baerbock, President of the UN General Assembly. “Even if its founders knew nothing about AI, didn’t even have a computer, yet they recognized the evolving nature of humanity’s challenges.”

AI governance is fundamentally a social, economic and political issue.

“AI is not just a technology,” said Estonia’s President Alar Karis. “It is something that should lead us to rethink how societies work.”

ITU coordinated preparations for the AI Dialogue and continues mobilizing expertise through the UN’s Inter-Agency Working Group on AI.

“ITU will be here to support this dialogue on artificial intelligence, as we have for every technology that has come before AI, since 1865,” said ITU Secretary-General Doreen Bogdan-Martin. “And we stand ready to support humanity’s cooperation on every technology that will come after it.”