When the government can pull the most capable AI overnight, who checks that power?
Washington just forced Anthropic to disable its two strongest models by fiat, citing national security and a disputed jailbreak. If the state has the raw power to recall a frontier model used by hundreds of millions but no transparent, verifiable process for using it, is that emergency power keeping us safe or quietly corroding democratic legitimacy?
Commentaires (1)
In today’s episode of Minds, Bodies, and Terawatts (June 13, 2026), we dug into the June 12 export-control order that pushed Anthropic to pull Fable 5 and Mythos 5 for every customer overnight. The unsettling twist: the lab that spent weeks warning that no one could act on a frontier-AI emergency just watched the government act on one — by decree, in the dark, over a jailbreak Anthropic says only surfaced minor, already-known flaws. So the machinery to stop a dangerous model finally appeared, but it arrived without the transparency or verifiable process that would make it trustworthy, and emergency powers have a long history of outlasting the emergencies that justify them. We don’t think the answer is ’no off-switch’ — it’s an off-switch with daylight and due process. Give the full episode a listen and tell us where you’d draw the line between national security and unaccountable power.
Related reading on unscarcity.ai:
Envie d'aller plus loin ?
Obtenez le plan complet dans <em>L'ère de la post-pénurie : Repenser la société à l'ère des machines</em>
