Fei-Fei Li Calls for Independent AI Safety Evaluation Standards
BIBIBI
AT A GLANCE
Fei-Fei Li said AI safety assessments should not be the sole responsibility of development companies, calling for academia, government, and industry to establish independent standards, and disclosed that World Labs is developing 3D world models for robot training.
Article
Fei-Fei Li: AI safety evaluations should not be the sole responsibility of developing companies; independent evaluation standards are needed
PANews reported on September 23, citing Bloomberg, that Fei-Fei Li, co-founder of World Labs and a Stanford University professor, stated that increasingly capable artificial intelligence systems should not have their safety evaluated solely by the companies developing them. Companies can internally test models' capabilities and safety properties, but academia, governments, and industry still need to jointly establish broader independent evaluation standards. She believes that AI's potential risks depend on how humans develop, govern, and use the technology, and that humans must retain control. Li also stated that World Labs is developing “world models” capable of generating and reconstructing 3D environments; the related technology can be used in scenarios such as robot training and evaluation.
Fei-Fei Li: AI safety evaluations should not be the sole responsibility of developing companies; independent evaluation standards are needed
02
PANews reported on September 23, citing Bloomberg, that World Labs co-founder and Stanford University professor Fei-Fei Li said increasingly capable artificial intelligence systems should not have their safety assessed solely by the companies developing them. Companies can internally test models' capabilities and safety attributes, but academia, government, and industry still need to jointly establish broader independent evaluation standards. She believes AI's potential risks depend on how humans develop, govern, and use the technology, and that humans must retain control. Li also said World Labs is developing “world models” capable of generating and reconstructing 3D environments, with the technology applicable to scenarios such as robot training and evaluation.
The following is analysis, separate from reported facts. Verify important claims independently.
Fei-Fei Li stated that AI safety assessments should not be conducted solely by development companies; academia, governments, and industry need to jointly establish independent standards. She also disclosed that World Labs is developing a three-dimensional world model for robot training.
Why it matters to readers
The news originates from Bloomberg and was republished by PANews, making the source reliable. The content reflects the public views of prominent AI scholar Fei-Fei Li, covering AI governance frameworks and World Labs' technological progress. It is relevant to AI industry regulatory trends, but is not a sudden market-impact event.
Follow subsequent official information and do not make decisions based on a single report.
Risks and unknowns
Whether regulatory agencies will respond to Fei-Fei Li's call and initiate work on developing independent AI assessment standards
Technical details or a product release from World Labs regarding its three-dimensional world model
How other AI giants will respond to independent assessment standards
