Paper: Team: @ChenShani2 @shi_weiyan @federicobianchy @Izzy_Gainsburg @jurafsky @RobbWiller . H/t to some of the many scholars whose work inspired ours: @baixx062 @jgvoelkel @JoshAGoldstein @JEichstaedt @KobiHackenburg @chris_bail .6/6.

These results imply that, while authorship labels likely enhance transparency, they are unlikely to substantially affect the persuasiveness of the labeled content, highlighting the need for alternative strategies to address challenges posed by AI-generated information. 5/6.

We also find that these patterns were not significantly different for respondents with different prior knowledge of the policy, political party affiliation, education level, age, or degrees of prior experience with AI. 4/6.

We find that authorship labels did not significantly influence participants’ attitude change toward the policy, judgments of the accuracy of the message, or their intentions to share the message with others. 3/6

We conducted a survey experiment (N=1601), presenting participants with an AI-generated message about several public policies, randomly assigning whether participants were told the message was generated by (a) an expert AI model, (b) a human policy expert, or (c) no label. 2/6

🚨🚨New Working Paper🚨🚨. AI-generated content is getting more politically persuasive. But does labeling it as AI-generated change its impact?🤔. Our research says the disclosure of AI authorship has little to no effect on the persuasiveness of AI-generated content. 🧵1/6