Not-AI, Therefore True? How AI Labeling Increases Information and Misinformation

• In Event: Lightning Talk: Information, Media, and Online Political Behavior

Brief Overview

AI labels are intended to fight deepfakes but increase belief in "cheapfakes" (miscaptioned real photos), while doing little to weaken beliefs in AI content. Users conflate authentic media with true claims.

Author

• Marli Wang Dunietz, Florida State University