Hinton, Ng, Li defend open AI despite safety concerns

The conference in Las Vegas brought together three influential voices in AI governance to defend openness against mounting industry concern. Geoffrey Hinton, a Nobel Prize winner, acknowledged the tension directly: open-weight models—trainable AI systems released to the public—are now permanent fixtures and carry real risks. "I was against open [weights] because it makes it so easy for people to take these big foundation models, which are very expensive to train, and for much less money train them to do bad things like cyber attacks," he explained. Yet Hinton recognized that the cost barrier has collapsed. "I think that battle's been lost. We now have open-weight models, so the barrier to lots of people getting these big models, which was the cost of training foundation models, that barrier has disappeared. It's too late." Despite his reservations, Hinton argued that AI's continued advancement would likely benefit society overall, boosting productivity and improving education and healthcare, and that concern about potential harms from smarter-than-human AI systems should not be dismissed as fearmongering.

Andrew Ng framed the issue through a different lens: competition and market dynamics. He voiced worry about gatekeepers limiting access to AI technology, comparing it to how Apple and Google control mobile operating systems. "I don't want there to be gatekeepers. That limits how all of us can access AI," Ng said, and his prescription was clear: "If I were to try to give one prescription, it would be to promote openness, because AI is amazing technology and I want it to be in everyone's hands." But Ng's defense of open models carried a geopolitical edge. He warned that if China's open-weight models gained widespread adoption across Asia, Africa, and the developing world, they could shape how billions of people encounter ideas about democracy, freedom, and human rights. "Whoever built the cheaper model would have the advantage," and if American open-source AI struggles to compete due to lobbying and fear-mongering, he cautioned, the cost-efficiency advantage would flow to Chinese systems.

Fei-Fei Li, CEO of World Labs, rejected both Hinton's caution and Ng's binary framing. "It's very dangerous to make this a dichotomy between complete openness all the way to complete closedness," she said. She drew parallels to nuclear physics, where scientific papers are published openly, uranium is strictly regulated, and laboratory work sits in between—demonstrating that different layers of knowledge management can coexist. Li also highlighted successes like the Human Genome Project, a collaboration between public and private institutions that created a platform enabling pharmaceutical companies to profit, scientists to advance their research, and society to benefit. "We need some levels of openness, both in scientific discovery, in education, in global partnership, as well as lucrative business models for entrepreneurs," Li argued. "But we also will accept closed-source systems."

Despite their tactical disagreements, all three speakers converged on one point: regulation would be necessary. Hinton was most explicit, saying "regulation will help us" keep AI development on track and that the decision cannot be left to tech leaders like Elon Musk and Mark Zuckerberg. Yet none of the three specified what that regulation should entail or how it would be enforced—a notable gap given the complexity of the technical and geopolitical challenges they each identified.