Persuasion Undermining Control

A threat-modeling framework covering AI communication that influences human decisions in ways which weaken the development, containment, oversight or governance of AI systems, paired with a decomposition of that risk into estimable variables.