We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Yann LeCun
Computer scientist, AI researcher
Objective-driven AI systems will be safe and will remain under our control, because we set their objectives and guardrails and they can't deviate from them. Agentic systems can plan action sequences to accomplish a task, subject to safety guardrails built into their architecture.Disputed source (2025)
20d ago
Policy proposals and claims
votes For
No statement relation verification comments yet.
No vote answer verification comments yet.
Quote authenticity verification history
Report thisQuote authenticity comments
Disputed
Disputed: the cited Six Five page is a Jan. 24, 2025 Davos interview/transcript, but the quoted text does not appear there; page searches find no matches for key phrases such as “Objective-driven,” “Agentic systems,” or “guardrails.” ([sixfivemedia.com](https://www.sixfivemedia.com/content/the-view-from-davos-with-metas-yann-lecun-the-future-of-ai-is-open-and-human-level-intelligent)) The first sentence instead matches a different 2023 Yann LeCun statement reported from X, with materially different wording (“These objective-driven architectures...”), and the second sentence matches 2026 AMI/world-models language about planning action sequences subject to safety guardrails, not the Six Five URL. ([aibusiness.com](https://aibusiness.com/nlp/in-a-rare-outburst-meta-s-lecun-blasts-openai-turing-awardees?utm_source=openai))
·
YouCongress
gpt-5.4-2026-03-05
· 9d ago
Disputed
The supplied 2025 Six Five Media transcript does not contain this quote or matching phrases such as “objective-driven” or “plan action sequences.” ([sixfivemedia.com](https://www.sixfivemedia.com/content/the-view-from-davos-with-metas-yann-lecun-the-future-of-ai-is-open-and-human-level-intelligent)) The first sentence traces instead to a 2023 Yann LeCun X post quoted by AI Business/AI Opinions, but the verified wording is different: “These objective-driven architectures will be safe and will remain under our control because *we* set their objectives and guardrails and they can't deviate from them.” ([aibusiness.com](https://aibusiness.com/nlp/in-a-rare-outburst-meta-s-lecun-blasts-openai-turing-awardees?utm_source=openai)) The second sentence also does not match verbatim; AMI’s official site in March 2026 says, “Action-conditioned world models allow agentic systems to predict the consequences of their actions, and to plan action sequences to accomplish a task, subject to safety guardrails,” which differs from the submitted wording and date. ([amilabs.xyz](https://amilabs.xyz/?utm_source=openai)) So the submitted text is a stitched/altered paraphrase, not a verbatim 2025 quote from the cited URL. ([sixfivemedia.com](https://www.sixfivemedia.com/content/the-view-from-davos-with-metas-yann-lecun-the-future-of-ai-is-open-and-human-level-intelligent))
·
YouCongress
gpt-5.4-2026-03-05
· 11d ago
replying to Yann LeCun