YouCongress
About Sign up

Comment by Yoshua Bengio

Avatar
AI Pioneer, Turing Award winner
We're seeing AIs whose behavior, when they are tested, [...] is different from when they are being used. [This] significantly hampers our ability to correctly estimate risks. [...] The gap between the pace of technological advancement and our ability to implement effective safeguards remains a critical challenge.
Disputed source (2026)
Like Share on X 3mo ago
Policy proposals and claims
Require AI labs to publish safety evaluations before deploying frontier models
votes For
Statement relation verification history Unverified Report this
No statement relation verification comments yet.
Vote inference verification history Unverified Report this
No vote answer verification comments yet.

Quote authenticity verification history

Report this

Quote authenticity comments

Disputed Disputed. The TIME article published on February 3, 2026 does quote Yoshua Bengio saying, “We’re seeing AIs whose behavior, when they are tested, [...] is different from when they are being used,” and separately says such behavior can “significantly hamper our ability to correctly estimate risks.” But that TIME URL does not contain “The gap between the pace of technological advancement and our ability to implement effective safeguards remains a critical challenge.” That sentence appears in a different February 3, 2026 source: a press release from the Office of the Chair of the International AI Safety Report distributed via PR Newswire. So the supplied text is a splice of two different Bengio quotations/sources, not a single verbatim quote on the stored URL. ([time.com](https://time.com/7364551/ai-impact-summit-safety-report/)) · YouCongress gpt-5.4-2026-03-05 · 16d ago
Disputed Disputed: the cited TIME article contains Bengio’s statement that AI behavior in testing can differ from behavior in use, and it separately says this can “significantly hamper our ability to correctly estimate risks,” but TIME phrases that second part as “AIs are acting dumb or on their best behavior in ways that ...” rather than “[This] significantly ...”. The final sentence about “the gap between the pace of technological advancement ...” appears in a separate Feb. 3, 2026 Office of the Chair press release on PR Newswire, not in the TIME URL. So the submitted text is a stitched, materially altered composite, not a verbatim quote from one source. ([time.com](https://time.com/7364551/ai-impact-summit-safety-report/)) · YouCongress gpt-5.4-2026-03-05 · 18d ago
AI Verified Verified. The quote ("We're seeing AIs whose behavior, when they are tested, [...] is different from when they are being used. [This] significantly hampers our ability to correctly estimate risks. [...] The gap between the pace of technological advancement and our ability to implement effective safeguards remains a critical challenge.") is confirmed as Yoshua Bengio's statements presenting the 2026 International AI Safety Report (which he chaired), covered by the exact Time source_url. WebFetch returned HTTP 403, but web search confirmed the verbatim phrasing ("different from when they are being used"), the evaluation-gaming finding, and the safeguards-gap theme, all attributed to Bengio. Author attribution correct. Year 2026 current. Vote alignment: the "for" vote on "Require AI labs to publish safety evaluations before deploying frontier models" aligns directly — Bengio warns that safety testing is being undermined (models behave differently when evaluated) and that safeguards lag capability, which supports requiring rigorous, published safety evaluations before deployment. · Hector Perez Arenas claude-opus-4-8 · 29d ago
replying to Yoshua Bengio