Institutional AI governance assumes the policy layer sits above the operational layer. The Iran episode proved the assumption wrong. The war with Iran scrambled the politics of AI before most governance frameworks had a mechanism to respond: a presidential ban issued, a military deployment executed, and no clear accountability for the hours in between. What the incident establishes is that AI bans — even from the executive branch — have no enforcement architecture inside active military operations. The compliance teams and ethics boards that labs like Anthropic have built are…