Imagine navigating an arena bustling with live coding, AI refresher courses, and a maze of cutting-edge gadgets. Attendees, often sporting glowing green headphones broadcasting UN panel discussions, weave through the exciting chaos. It’s a vibrant, sometimes overwhelming, scene that defines the unique atmosphere of the AI for Good Summit.
Amidst this futuristic backdrop, you might find yourself in the Networking Zone, perhaps even trying to engage with others on a rotating seating contraption dubbed UFOTECH. While it might resemble a giant lazy Susan more than a serious networking bench, it perfectly encapsulates the summit’s blend of innovation and unconventional approaches to global challenges.
Focusing on “AI for Good”
This remarkable event, now in its 10th year, is orchestrated by the United Nations’ International Telecommunication Union (ITU). Its core mission is ambitious: to convene global leaders from both the private and public sectors, urging them to collaboratively harness artificial intelligence for humanity’s benefit, not its detriment.
This focus notably diverges from the often alarmist discussions dominating headlines, where Silicon Valley executives and AI lab leaders testify to lawmakers about superintelligence risks. While export controls on chips and existential fears capture Washington’s attention, the UN AI for Good Summit consistently champions a more idealistic, solution-oriented vision for AI’s future.
Doreen Bogdan-Martin, Secretary-General of the ITU, articulated this conviction in her keynote. She emphasized the belief that “artificial intelligence, deployed responsibly, could help solve humanity’s most pressing problems—from hunger to disease to a warming planet.” This foundational idea, she noted, is being rigorously tested by the very challenges AI itself presents.
The Elusive Definition of “Good”
The vast 106,000-square-meter convention center on the outskirts of Geneva’s airport district buzzed with a pervasive question: what truly constitutes “good” when applied to AI, and what tangible benefits does it deliver? Sessions were underpinned by a growing apprehension that unchecked deployment by powerful corporate monopolies is already hardwiring global inequality and eroding fundamental human rights.
For those on the front lines, the tech industry’s utopian veneer has begun to wear thin. Giulio Coppi, a senior humanitarian officer with campaign group Access Now, voiced concerns about the public and humanitarian sectors’ excessive reliance on big tech. He stressed that we should be “out of the age of innocence,” urging organizations to stop treating tech companies “as your best friends.”
Coppi highlighted a decade of opaque, multi-million dollar deals funded by public money, questioning the lack of transparency. He warned that many organizations can’t even explain the components of their tech stack because it’s constantly changing and controlled by external entities. This skepticism was further amplified by dramatic events, including pro-Palestine activists interrupting a keynote by Amazon CTO Werner Vogels, alleging the company’s technology aids Israeli forces.
Harvard engineering professor Vijay Janapa Reddi also pointed out a fundamental challenge: the vagueness of “good” as an engineering standard. Amidst the din of competing sessions, he observed, “When you’re an engineer, good means nothing. I can’t build you something that is good. A plane that flies for five minutes ain’t no good.” This highlights the difficulty in translating high-level ethical aspirations into practical, measurable outcomes for AI development.
Bridging the AI Divide: Access and Infrastructure
Much of the contemporary global debate surrounding AI now centers on access: who has the ability to use advanced models, purchase essential chips, and participate in the burgeoning compute economy. This crucial issue underscores decisions like the Trump administration’s implementation and subsequent lifting of export controls on frontier AI models, and China’s reported deliberations on making its open-weight models less accessible.
Restricting access and excluding poorer countries risks making them perpetually dependent on foreign infrastructure, platforms, and technical standards. In a vital session on AI hardware and the widening digital divide, speakers underscored that compute power is no longer merely a technological concern, but a critical development challenge. Syed Munir Khasru, chairman of the Institute for Policy, Advocacy, and Governance, argued, “If we mean AI for good, meaning compute for all, we must recognize that this is [about] development infrastructure, not just technology.”
Speakers also highlighted that most large language models (LLMs) are currently structured around English, creating significant barriers for non-English speaking communities. The solution, they suggested, lies in developing smaller, local LLMs that can run efficiently on more affordable hardware. This would ensure AI serves a broader global population, extending beyond the wealthiest markets and fostering true inclusivity.
The politics of infrastructure—determining who receives what, and who maintains control—was a constant undercurrent throughout the summit. The core question transcended mere safety; it pondered whether nations outside the established US-China-Europe axis would truly be empowered to shape the trajectory and applications of this transformative technology.
From Talk to Action: Charting a Responsible Future
Traditionally, discussions around technical standards have rarely involved considerations of end-user rights or representation, as Gilles Thonet, deputy secretary-general of the International Electrotechnical Commission, observed. He told WIRED that engineers might consider human rights “someone else’s business,” but emphasized that, in fact, “they’re not.” Anja Kaspersen, director of global markets development at IEEE, further highlighted that the most consequential decisions often aren’t made in UN assemblies, but are embedded within hidden architectures, technical standards, and procurement choices.
To address this, Kaspersen advocates for building “middleware”—a connective layer that can translate abstract high-level human rights principles into verifiable, technical enforcement mechanisms. Jeremy Ng, counsel for AI and the digital economy at the World Bank, reinforced this, stressing that AI impact assessments must evolve into practical tools with real teeth, moving beyond mere “governance theater” or superficial box-ticking exercises for tech giants.
While the summit saw ample discussion, concrete action is often the next, more challenging step, requiring significant consensus. The UN, however, celebrated a pivotal advancement: the formation of a 44-member commission specifically designed to shepherd in AI for Good. This crucial body will be co-chaired by Rwandan President Paul Kagame and Salesforce CEO Marc Benioff, signifying a powerful collaboration between governmental and industry leaders.
As Bogdan-Martin stated in her opening keynote, “No single stakeholder can shape the future of AI alone. It needs builders. It needs you.” Yet, even amidst this call for collaboration, a striking visual on the bustling convention floor underscored the persistent tension. Tesla Cybertrucks were showcased alongside UN rescue helicopters, and humanoid robots darted rapidly between booths, drawing stares from human attendees. This undeniable speed of technological advancement often outpaces the measured, deliberate process of forging global consensus on what “good” truly means for AI.
Source: Wired – AI
