World Labs Founder Fei-Fei Li Urges Independent Oversight of AI

Artificial intelligence pioneer Fei-Fei Li called for more oversight of the technology by independent and public-sector entities, arguing that safety assessments of increasingly capable AI systems shouldn’t be left solely to the companies developing them.

Li, a Stanford University professor and co-founder of World Labs, said Tuesday that while researchers inside AI companies can measure the power and security attributes of individual models, that’s no substitute for broader standards developed by governments, industry and academic institutions.

“I continue to believe in the importance of benchmarks from independent bodies and public sector bodies like academia as well as the shared responsibility of industry, of government together,” Li said in a Bloomberg Television interview.

Anxiety about the potential risks posed by increasingly powerful AI systems has surged following warnings from Silicon Valley leaders that the technology is advancing so fast that it may extend beyond human control. In recent weeks, Anthropic PBC Chief Executive Officer Dario Amodei and OpenAI CEO Sam Altman have called for a slower pace of development for cutting-edge models. Both men are set to discuss AI safety with the United Nations Security Council on Wednesday.

Li said that AI’s potential existential risks ultimately depend on how people choose to develop, govern and use the technology. This principle applies beyond AI, Li said, arguing that technologies from electricity and steam engines to modern transportation have similarly required society to balance potential benefits with human safety.

Read More: Trump Rejects Push for International Pact to Manage Risks of AI

Humans “own the power of technology,” Li said. “We must be in control,” she said. “AI is not about AI, it is about humans. It is about our shared as well as individual dignity and agency, as well as our responsibility.”

Li’s comments add to an intensifying debate among policymakers and AI industry leaders over how seriously to take the warnings about technology’s most extreme risks. Earlier Tuesday, US President Donald Trump rejected concerns over AI’s existential threat, calling them “a hoax” during remarks to the UN General Assembly in New York and stressing instead the need to continue developing the technology at full speed.

Read More: AI Pioneer Andrew Ng Calls Extinction Fears ‘Science Fiction’

Discussions over how AI systems behave is extending beyond the so-called frontier models driving the current boom, Li said. Her World Labs venture is developing so-called world models that can generate and reconstruct three-dimensional environments, technology that she said could be used to train and evaluate robots as well as power applications in design, healthcare and other industries.

As those systems move toward real-world applications, Li said the industry faces the same fundamental question: how to evaluate increasingly capable AI and determine whether it is behaving as intended.