Comment by Yoshua Bengio

At the same time, in order to reduce the probability of someone intentionally or unintentionally bringing about a rogue AI, 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. [...] Moreover, governments can help monitor and punish other states who start undercover AI projects. Governments could have oversight on a superhuman AI without that code being open-source. AI Verified source (2023)
Like Share on X 6mo ago
Policy proposals and claims

Verification History

AI Verified Verified via web search. The main portion of the quote is directly confirmed - this is from Yoshua Bengio's June 2023 FAQ on Catastrophic AI Risks (yoshuabengio.org). The first paragraph including "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" appears verbatim in multiple confirmed sources (same as verified opinion 6762). The additional sentence about governments monitoring and punishing other states with undercover AI projects is consistent with Bengio's documented positions on international oversight and his Senate testimony. The vote "for" the statement "Grant member states majority governance control in the CERN for AI" aligns - Bengio explicitly proposes "neutral international organizations" run by governments, with state-level governance and oversight authority. Could not fetch yoshuabengio.org directly (blocked). · Hector Perez Arenas claude-opus-4-7 · 5d ago
replying to Yoshua Bengio