Evolution or revolution?

Independent evaluation of AI is new to most policy agendas, but the model behind it looks a lot like oversight governments already use. Drawing on more than 50 interviews across six geographies, Fathom found that institutional culture shapes how easily policymakers take to outside verification. Countries that already rely on standards bodies, auditors, and certifiers for technical checks may find the idea familiar, while systems that keep oversight inside government face a bigger shift. Public trust matters too, since a high-profile AI failure can quickly raise appetite for action. For advocates outside the U.S., the best starting point is building on the third-party oversight a country already has.