
Could AI Manipulate an Entire Election Without Anyone Noticing?
Could an election be counted correctly while the information shaping voters’ choices was quietly distorted? This episode examines AI election manipulation beyond the familiar fear of a convincing deepfake. We look at experiments in political persuasion, unreliable chatbot answers and the difference between producing propaganda and reaching people who believe it.
The central challenge is evidence. A changed opinion in a study is not a changed ballot, and a detected influence operation does not prove that an election’s winner would otherwise have been different. Explore why private AI…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.