Dodge beyond the live coding sessions on stage, AI refresher courses, a gadget obstacle course, round people walking around with bright green silent disco-style headphones blasting UN panel discussions in your ears, and you can pause for a breath. But you might find yourself in the Networking Zone, in a swivel-seating contraption called UFOTECH that looks more like the kind of lazy Susan you’d find in a Chinese restaurant than the networking bench it’s designed for.

This is the AI ​​for Good Summit, organized by the United Nations’ International Telecommunication Union (ITU), where representatives from the public and private sectors attempt to discuss how to harness technology to the benefit, rather than the detriment, of humanity.

As Silicon Valley executives and AI lab leaders testify before lawmakers in Washington about the risks of superintelligence, and the White House imposes controls on chip exports, the UN’s AI for Good Summit, now in its 10th year, is focused on much more idealistic goals.

“Our belief that artificial intelligence, implemented responsibly, could help solve humanity’s most pressing problems, from hunger to disease to global warming,” said Doreen Bogdan-Martin, ITU Secretary-General, in a keynote address on the conference’s main stage. “Today, that idea is being tested, even by the challenges that AI itself poses, even as we strive to use it for good.”

What good means (and what good it does for humanity) was a divisive question throughout the conference, which took place in a massive 106,000-square-meter convention center on the edge of Geneva’s airport district. The sessions were underpinned by deep concern that lackadaisical deployment by unbridled corporate monopolies is already generating global inequality and eroding human rights.

For some on the front lines, the utopian veneer of the tech industry has already worn away. Speaking on the sidelines of the event, Giulio Coppi, a senior humanitarian official at campaign group Access Now, denounced the public and humanitarian sectors’ over-reliance on big tech. “We should be past the era of innocence,” says Coppi, demanding that organizations stop treating technology companies “like their best friends.” He points to a decade of opaque, multi-million dollar deals financed with public money. “You can’t even explain what’s inside their technology stack, because it has continued to change,” he warns.

The United Nations wants to treat AI with the same urgency as climate change

Coppi’s opposition was muted compared to some: pro-Palestinian activists stormed the stage during a keynote speech by Amazon’s chief technology officer Werner Vogels, alleging that Israel is using the company’s technology against Palestinians, before eventually being kicked off the premises.

“When we talk about AI, we love the buzz, we get excited,” says Vijay Janapa Reddi, an engineering professor at Harvard University, of the din of competitive sessions during a presentation. “The damn thing never comes into practice.” The problem, he says, is that “good” is too vague a standard to counteract. “When you’re an engineer, good doesn’t mean anything. I can’t build you something that’s good. A plane that flies for five minutes isn’t good.”

Much of the global debate around AI now focuses on access: who can use the models, who can buy the chips, and who is excluded from the computing economy. It’s part of the reason the Trump administration implemented, and then removed, export controls on major frontier AI models, and China is reportedly considering making its open models less open. Restricting access and excluding poorer countries can leave them dependent on foreign infrastructure platforms and standards.

In a session on AI hardware and the growing digital divide, speakers argued that computing is no longer simply a technological problem, but a development problem. “If we want to say AI for good, that is, computing for all, we must recognize that this is [about] development infrastructure, not just technology,” says Syed Munir Khasru, president of the Institute of Politics, Defense and Governance. Others noted that most large language models remain structured around English, making smaller local LLMs running on cheaper hardware essential for AI to serve communities beyond the wealthiest markets.