Who should decide what an AI refuses to do: its makers, its buyers, or the state?
If a government can label an AI vendor a national security risk because its model refuses certain tasks, does that make encoded ethical limits a public policy question rather than a company’s choice? And if so, who gets the final say over an AI’s hard lines: the founders who wrote them, the customers who pay for the tool, or the state that wants it to comply?
Commentaires (1)
In today’s episode of Minds, Bodies, and Terawatts, dated September 26, 2026, the hosts dig into the DC Circuit’s 2-1 ruling upholding the Pentagon’s decision to designate Anthropic a supply chain risk over Claude’s refusals on autonomous weapons and mass surveillance of Americans. The episode weighs the majority’s practical argument that a tool which might quit mid-mission is a real operational hazard against the dissent’s warning that any contractor can now be told to change its AI-use policy or be branded a threat. It also connects the ruling to the same week’s news that Anthropic’s founders want to keep control of those limits after the IPO, and asks whether an AI’s conscience belongs to its creators, its owners, or the government. Listen to the full episode, then tell us where you would draw the line and who should hold the pen.
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>
