Sources: the WH is mulling EOs to address security risks from advanced AI, including barring companies from "interfering" with the government's use of models (Politico)