Fei-Fei Li on Require frontier AI labs to conduct nea…

Fei-Fei Li

Stanford AI professor; HAI co-director

It’s very dangerous to make this a dichotomy between complete openness all the way to complete closedness. In complex software systems as well as scientific systems it’s much more nuanced.

AI Verified (Aug 12, 2026)

Source:

Fei-Fei Li, at the Ai4 conference, as reported by TechCrunch, August 12, 2026.

2mo ago

Policy proposals and claims

Require frontier AI labs to conduct nearly all R&D publicly but keep the weights closed

votes For

Statement relation verification history AI Verified Report this

AI Verified Relevant: the quote directly addresses the openness-vs.-closedness tradeoff in frontier AI, and the source context says Li argues openness should vary by layer rather than be all-or-nothing, which makes her nuanced stance on a mixed public-R&D / closed-weights policy determinable. ([tech.yahoo.com](https://tech.yahoo.com/ai/articles/ai-safety-concerns-mount-three-175100465.html?utm_source=openai)) · YouCongress gpt-5.4-mini-2026-03-17 · 1mo ago

AI Unverifiable Li advocates context-dependent degrees of openness but does not determine a stance on this specific public-R&D/closed-weights requirement. · Hector Perez Arenas gpt-5.6 · 1mo ago

Vote inference verification history AI Verified Report this

AI Verified For: Li said it’s dangerous to frame AI as a choice between “complete openness” and “complete closedness,” and the article context describes her as favoring a more nuanced public-private model, citing the Human Genome Project. That makes a hybrid policy with public R&D and closed weights the closest fit, though she did not endorse this exact wording. ([tech.yahoo.com](https://tech.yahoo.com/ai/articles/ai-safety-concerns-mount-three-175100465.html?utm_source=openai)) · YouCongress gpt-5.4-mini-2026-03-17 · 1mo ago

Quote authenticity verification history

Report this

AI Verified The TechCrunch report of the Ai4 conference reproduces Fei-Fei Li's wording verbatim. · Hector Perez Arenas gpt-5.6 · 1mo ago

replying to Fei-Fei Li

Other opinions from Fei-Fei Li

See all

Fei-Fei Li votes Against

AI companies should be liable for harms caused by their deployed models

SB-1047 holds liable [...] the original developer of that model. It is impossible [...] to predict every possible use. SB-1047 will force developers to pull back [...].

AI Verified source (2024)