Culture & Human Agency

A warning helped people resist AI political persuasion

A study of 3,208 Americans found that a short warning about persuasive AI cut belief change by about 48%. The warning did not just say an AI was involved. It told people that language models can be prompted to persuade and show selective information.

AI Agency
Read original source

What the source reports

The paper reports a study with 3,208 Americans in political conversations with language models. A language model is an AI model that works with words. A brief warning cut belief change by about 48%. The snapshot does not say which topics were tested or how long the effect lasted.

Original source

Title
A light-touch AI literacy intervention helps protect against AI political persuasion
Author
Reed Orchinik, David Rand
Publication
arXiv
Date
Monday, September 14, 2026