The US government's secretive lifting of restrictions on Anthropic's AI models, following an improvised ban, unnerves global users. This opacity regarding safety agreements and future protocols fuels uncertainty and favoritism concerns. The author criticizes the lack of clear, public standards, advocating for transparent AI governance. They suggest outsourcing security testing to better-funded, neutral agencies like the UK's AI Security Institute. Public standards are crucial to prevent "kill-switch" risks, maintain international trust in American AI, and avert market instability.