Comment by Yoshua Bengio

we need to increase governance and we should consider limiting access to the large-scale generalist AI systems that could be weaponized, which would mean that the code and neural net parameters would not be shared in open-source and some of the important engineering tricks to make them work would not be shared either. Ideally this would stay in the hands of neutral international organizations (think of a combination of IAEA and CERN for AI) that develop safe and beneficial AI systems that could also help us fight rogue AIs.
Verified source
Like Share on X 10mo ago
Policy proposals and claims
votes For
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified The quote clearly supports restricting open-source release of dangerous AI: it says systems that "could be weaponized" should have their code and weights "not be shared in open-source." That strongly implies support for banning open-source AI models capable of creating WMDs, even if phrased as "should consider" limiting access. · YouCongress gpt-5.4-2026-03-05 · 18d ago
Vote inference verification history AI Unverifiable Report this

Vote answer comments

AI Unverifiable The quote says we "should consider limiting access" and that such systems "would not be shared in open-source," but this is framed as a consideration and applies to systems that "could be weaponized," so it does not unambiguously endorse the full ban statement. · YouCongress gpt-5.4-2026-03-05 · 18d ago
votes For
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified The quote clearly supports the idea of a global AI institution: it says AI should ideally stay in the hands of 'neutral international organizations' and explicitly references 'a combination of IAEA and CERN for AI.' That directly implies support for creating an international CERN-like AI body. · YouCongress gpt-5.4-2026-03-05 · 18d ago
Vote inference verification history Latest opinion AI Verified Report this

Vote answer comments

AI Verified The quote clearly supports this: it says AI should "stay in the hands of neutral international organizations" and explicitly says to "think of a combination of IAEA and CERN for AI." · YouCongress gpt-5.4-2026-03-05 · 18d ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Verified. The source page is credited to Yoshua Bengio and published on June 24, 2023; it contains the quoted text verbatim across consecutive lines 187–188, including the sentences about limiting access, not sharing code/parameters in open source, and "neutral international organizations" like "IAEA and CERN for AI." ([yoshuabengio.org](https://yoshuabengio.org/2023/06/24/faq-on-catastrophic-ai-risks/)) · YouCongress gpt-5.4-2026-03-05 · 19d ago
Verified Migrated from legacy verification · Hector Perez Arenas · 9mo ago
replying to Yoshua Bengio