Robot Canines, Electric Cars, and Emergency Choppers: The UN AI Summit Had It All

Navigate through the live coding demonstrations, AI review workshops, a maze of gadgets, individuals meandering about with illuminated green silent-disco-style headphones broadcasting UN panel discussions into your ears, and you can take a moment to catch your breath. However, you may find yourself in the Networking Zone, on a rotating seating apparatus known as UFOTECH, resembling the kind of lazy Susan you’d find at a Chinese restaurant more than the networking bench it was intended to be.

This is the AI for Good summit, organized by the United Nations’ International Telecommunication Union (ITU), where delegates from both private and public sectors convene to explore how to leverage technology for humanity’s benefit rather than its detriment.

While executives from Silicon Valley and leaders of AI labs are testifying to lawmakers in Washington about the dangers of superintelligence, and the White House imposes export controls on chips, the UN AI for Good Summit—now celebrating its 10th year—aims at more optimistic objectives.

“We are firmly convinced that responsibly deployed artificial intelligence can help tackle humanity’s most urgent challenges—from hunger and disease to climate change,” stated Doreen Bogdan-Martin, secretary-general of the ITU, during her keynote address on the main stage. “This vision is currently being tested, including the hurdles posed by AI itself, as we endeavor to utilize it for positive outcomes.”

The definition of good—and its implications for humanity—was a recurring theme throughout the conference held in an expansive 106,000-square-meter convention center located on the outskirts of Geneva’s airport district. Sessions were accompanied by a persistent concern that careless implementation by unchecked corporate giants is already embedding global inequality and diminishing human rights.

For many working at the forefront, the tech industry’s idealistic facade has already faded. Giulio Coppi, senior humanitarian officer at campaign group Access Now, criticized the humanitarian and public sectors’ excessive dependence on major tech companies. “We should have moved past the age of naivety,” Coppi asserts, calling for organizations to stop regarding tech firms “as your best allies.” He highlights a decade’s worth of unclear, multimillion-dollar contracts funded by public resources. “You’re unable to explain what’s in your tech stack, because it constantly changes,” he cautions.

The United Nations Wants to Treat AI With the Same Urgency as Climate Change

Coppi’s dissent was subdued in comparison to others: Pro-Palestine activists interrupted a keynote by Amazon’s chief technology officer, Werner Vogels, claiming the company’s technology is utilized by Israel against Palestinians before being escorted out of the venue.

“When discussing AI, we often bask in the hype and feel excited,” remarked Vijay Janapa Reddi, an engineering professor at Harvard University, amidst the cacophony of competing sessions during his presentation. “The reality is that it rarely translates into practice.” He emphasizes that “good” is an ambiguous criteria for engineering. “As an engineer, the term good is meaningless. I can’t create something that is good. A plane that only flies for five minutes isn’t valuable.”

The ongoing global dialogue surrounding AI is increasingly framed in terms of access: who can utilize the models, who can purchase the chips, and who is excluded from the computational economy. This is part of the rationale behind the Trump administration’s imposition, and subsequent removal, of export controls on leading frontier AI models, while reports suggest China is considering reducing the openness of its models. Restricting access and marginalizing poorer nations could make them reliant on foreign infrastructure and standards.

During a session focused on AI hardware and the expanding digital divide, speakers contended that computation has evolved into more than just a technological issue; it is a developmental challenge. “If we’re advocating for AI for good, which translates to compute for all, we must recognize that this pertains to developmental infrastructure, not simply technology,” states Syed Munir Khasru, chairman of the Institute for Policy, Advocacy, and Governance. Others noted that most large language models are predominantly focused on English, emphasizing the necessity for smaller, localized LLMs operating on more affordable hardware if AI is to meet the needs of communities outside the wealthiest markets.