Fei-Fei Li Says Independent Oversight Needed for AI Safety Assessments

Fei-Fei Li, a Stanford professor and co-founder of World Labs, said on Tuesday that independent institutions and the public sector should play a stronger role in overseeing AI. She argued that safety assessments of increasingly capable AI systems should not be left entirely to the companies that develop them.

Speaking in a Bloomberg Television interview, Li said internal researchers at AI companies can evaluate individual models, but that cannot replace broader standards established by government, industry and academia. She stressed that it is important for independent institutions and public bodies such as academia to set benchmarks, while industry and government also share responsibility.

Li noted that the existential risks associated with AI ultimately depend on how humans choose to develop, govern and use the technology. She drew parallels with previous technologies, saying society has always needed to balance benefits against physical safety.

She also mentioned that World Labs, the company she co-founded, is developing world models capable of generating and reconstructing three-dimensional environments. The technology could be used to train and evaluate robots, as well as support applications in design, healthcare and other industries, she said.