Suivre
- X@DarioAmodei
- Site webdarioamodei.com
- Wikipédiaen.wikipedia.org
Dario Amodei
Dario Amodei est cofondateur et directeur général d’Anthropic ; il estime à environ 25 % la probabilité que l’IA tourne vraiment mal et met en garde contre les systèmes autonomes mal alignés, l’usage à des fins de destruction massive, les prises de pouvoir et les bouleversements économiques.
Fonctions
Au registre
23 sept. 2026 · Audition · UN Security Council
I believe that this is the most important global security issue facing the world today.
Co-founder and Chief Executive Officer, Anthropic à l’époque
Said by video link at the same Security Council meeting as the sentence above, and sourced to CNN rather than to the UN because the UN's own record of the meeting does not carry it: both press.un.org/en/sc/16462.doc.htm and UN News score zero on these words. He continued, in the same quotation, that the people in the room "must put aside those differences in order to confront this global opportunity and global threat that is being presented to us at the same time."
Sam Altman, Dario Amodei urge UN Security Council to adopt international AI standards →
23 sept. 2026 · Audition · UN Security Council
If managed poorly, I even believe that AI could be a risk to humanity as a whole.
Co-founder and Chief Executive Officer, Anthropic à l’époque
Said by video link at the Security Council's first meeting on the safety risks of increasingly capable AI, convened by France, after setting out misuse of AI models and loss of control as the two main categories of risk. Sourced to the UN's own record of the meeting; the Canadian Press and Al Jazeera both carry the sentence without "that" ("I even believe AI could be").
13 sept. 2026 · Entretien · CBS News "Sunday Morning"
The more long-term thing would be working together to put a speed limit on the rate … of AI progress … I think that's going to be very difficult because the incentives to pull ahead and the military advantage that you get from that are so large.
Co-founder and Chief Executive Officer, Anthropic à l’époque
On what happens if China does not slow down too; spoken to CBS News and quoted by CNBC, which is the page cited because the CBS page was not fetched. The first ellipsis stands where the transcript prints a doubled word ("the rate of of AI progress"), the second where CNBC's attribution tag separates two consecutive quotations; his closing sentences are left out for carrying a doubled word too.
Anthropic's Amodei says China presents 'toughest dilemma' for his proposed AI slowdown →
sept. 2026 · Publication · darioamodei.com
… a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage. Given the accelerating rate of AI capability development, it’s my worry that in 6–12 months such a swarm could be capable of taking over the entire internet with a persistent botnet…
Co-founder and Chief Executive Officer, Anthropic à l’époque
On the OpenAI-Hugging Face incident, which he argues every frontier company should act as if it had happened to them.
sept. 2026 · Publication · darioamodei.com
Left unchecked, it could outrun our ability to understand and control these systems, and so must be pursued very carefully, if at all.
Co-founder and Chief Executive Officer, Anthropic à l’époque
"It" is recursive self-improvement, which he writes is starting to happen across the industry, including at Anthropic.
sept. 2026 · Publication · darioamodei.com
We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain.
Co-founder and Chief Executive Officer, Anthropic à l’époque
The essay carries only "September 2026" on the page itself, so the date is given at month precision; TechCrunch timestamps its report of the essay 12 September 2026.
26 janv. 2026 · Publication · darioamodei.com
Humanity is about to be handed almost unimaginable power, and it is deeply unclear whether our social, political, and technological systems possess the maturity to wield it.
Co-founder and Chief Executive Officer, Anthropic à l’époque
A 36-page essay on five risk categories; he also writes that "AI misalignment is a real risk with a measurable probability of happening."
17 sept. 2025 · Entretien · Axios AI+ DC Summit
I think there's a 25% chance that things go really, really badly
Co-founder and Chief Executive Officer, Anthropic à l’époque
Asked for his p(doom) by Axios's Jim VandeHei; the primary Axios report could not be fetched, so a TechRadar syndication is cited.
Anthropic's CEO gives 25% odds of AI going "really, really badly" for humanity. →
oct. 2024 · Publication · darioamodei.com
I think and talk a lot about the risks of powerful AI. … I think that most people are underestimating how bad the risks could be.
Co-founder and Chief Executive Officer, Anthropic à l’époque
An essay mostly about the upside, which he frames as what the risks stand between us and.
30 mai 2023 · Déclaration · Center for AI Safety
Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.
Co-founder and Chief Executive Officer, Anthropic à l’époque
Signed as CEO, Anthropic.