The United Nations-backed AI for Good Global Summit opened on Tuesday in Geneva, bringing together governments, technology firms, academics and civil-society groups to discuss how artificial intelligence can advance the Sustainable Development Goals while addressing mounting safety and governance concerns. The four-day event, organised by the International Telecommunication Union in partnership with more than 40 UN agencies and co-hosted by the Swiss government, runs from 8–11 July. ITU Secretary-General Doreen Bogdan-Martin told delegates that rapidly advancing “agentic” AI systems demand urgent, inclusive oversight. Citing an ITU survey showing that 85% of countries still lack an AI-specific policy, she warned that the greatest risk lies in deploying the technology without fully understanding its societal and environmental implications. More than 20,000 square metres of exhibition space at Geneva’s Palexpo feature over 200 demonstrations, ranging from flying cars and brain-computer interfaces to water-quality monitors and AI-powered disaster-response tools. A dedicated health track led by the World Health Organization on Wednesday will examine how AI can improve diagnostics and emergency care in low-resource settings and preview WHO guidance on traditional-medicine applications. The agenda also includes an AI Governance Day on Thursday, where regulators and international bodies will explore standards and cross-border cooperation. The summit will conclude with the AI for Good Awards, recognising projects that apply AI for public benefit across people, planet and prosperity.