The Trump administration's framework for assessing potential cybersecurity risks posed by advanced AI reportedly has no interest in testing open models. Axios reports that not only do the voluntary guidelines outright exclude open models - meaning anyone can download them and inspect their core comp
