À qui la faute ? À Anthropic, à l'administration Trump ? À l'intelligence artificielle ? Qui prend la responsabilité ?
Links
📺 Website / YouTube
https://www.youtube.com/@Parlons.IA.2
https://www.youtube.com/@ParlonIA
🐦 Twitter / X
https://x.com/ParlonsIAx
🎧 Podcast
https://open.spotify.com/show/1ThhxveRkTiSGR09cZsrPR
📩 Direct contact
formation.ai87@gmail.com
How Anthropic got tied to the killing of children in Iran
Here is the part that matters. Anthropic did not appear from nowhere. The company had already secured a Department of Defense agreement with a ceiling of $200 million in July 2025 to prototype frontier AI capabilities for U.S. national security. It also publicly announced Claude Gov models built for classified U.S. national security environments. This was not a toy deployment. This was the normalization of Anthropic inside serious state power.
Then came Minab. Public reporting says the Pentagon used Anthropic’s Claude in planning operations against Iran, and there are now mounting questions about whether AI-assisted targeting, outdated intelligence, or both helped produce the strike on the school. A preliminary U.S. military investigation, according to current reporting, says the U.S. was likely responsible for the strike. That still does not prove Anthropic directly selected the school as a target. But it absolutely does mean the company’s technology is now entangled with one of the most horrifying civilian attacks of this war.
That distinction matters. If you say more than the evidence supports, you weaken the argument. If you say less, you sanitize reality. The honest version is brutal enough: Anthropic built and supplied systems for U.S. national security; those systems were reportedly used in military planning and targeting workflows during the Iran campaign; a girls’ school was struck; and the public still does not have a full, transparent account of how the decision chain worked. That is already damning.
