Dodge past live on-stage coding sessions, AI refresher courses, an obstacle course of gizmos, people circling around with disco-style glittery blue headphones blasting UN panel discussions into your ears, and you can take a break. But you might find yourself in the Networking Zone, in a revolving seating area called UFOTECH that looks more like the kind of lazy Susan you’d meet at a Chinese restaurant than a networking bench designed to function like.
This is the AI for Good conference, organized by the United Nations Telecommunications Union (ITU), where representatives of the private and public sectors try to discuss how they can use technology for the benefit, rather than harm, of humanity.
While Silicon Valley executives and AI lab leaders testify to lawmakers in Washington about the dangers of artificial intelligence, and the White House slaps export controls on chips, the UN AI for Good Summit—now in its 10th year—focuses on more sensible goals.
“It is our belief that artificial intelligence, used responsibly, can help solve the problems most pressing humanity—from hunger to disease to global warming,” said Doreen Bogdan-Martin, ITU secretary-general, in a keynote speech on the conference’s main stage. “Today, that vision is being tested, including the challenges that AI brings, as we strive to use it for good.”
What are good definitions—and what good do they do for humanity—was the question asked throughout the conference, which spread across a massive 106,000-square-meter convention center on the edge of Geneva’s airport district. Sessions is backed by a drumbeat of concern that reckless exports by unscrutinized private companies are already exacerbating global inequality and undermining human rights.
For some frontrunners, the utopian veneer of the tech industry is already wearing thin. Speaking on the sidelines of the event, Giulio Coppi, chief humanitarian officer at the campaign group Access Now, denounced the over-reliance of aid companies and the public on big technology. “We have to get out of the age of innocence,” Coppi said, demanding that organizations stop treating technology companies as “your best friends.” He points to a decade of vague, multimillion-dollar deals funded by public money. “You can’t even define what’s inside your technology stack, because it’s constantly changing,” he warns.
Coppi’s opposition was muted compared to others: Pro-Palestinian activists stormed the stage during a keynote speech by Amazon’s chief technology officer, Werner Vogels, complaining that the company’s technology was being used by Israel against the Palestinians, before it was eventually removed from the site.
“When we talk about AI, we love the hype, we enjoy it,” said Vijay Janapa Reddi, an engineering professor at Harvard University, during a competitive pitch during the launch. “The damn thing actually never worked.” The problem, he says, is that “good” is too vague a level that an engineer can handle. “If you’re an engineer, good doesn’t mean anything. I can’t build something good for you. A plane that flies for five minutes is not good.”
Much of the global debate about AI is now framed around 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 why the Trump administration has implemented, and then lifted, export controls on advanced AI models across the border, and China is reportedly considering making its open-source models unopened. Strengthening access and cutting off the poorest countries can leave them dependent on external infrastructure and standards.
In a session on AI hardware and the expanding digital divide, speakers argued that computing is no longer just a technology problem, but a development problem. “If we mean for AI to be good, which means that it includes everyone, we must realize that this is the case [about] infrastructure development, not just technology,” said Syed Munir Khasru, chairman of the Institute for Policy, Advocacy, and Governance.” Others have pointed out that most of the major language models are still programmed in English, making local LLMs using cheap hardware essential if AI is to serve communities beyond the wealthiest markets.
