Towards scalable AI-assisted pre-bunking of election misinformation: evidence from a pre-registered US panel experiment

Abstract

Abstract We present a scalable artificial intelligence (AI)-assisted framework for rapidly generating ‘pre-bunking’ articles against election misinformation. Our approach uses a stable template structure that, once finalized through human review, can be quickly adapted to counter emerging false narratives. In a pre-registered two-wave experiment with 4293 United States (US) registered voters, we test this framework against politically charged election misinformation—one of the most challenging domains for misinformation intervention. We find that large language model (LLM)-generated pre-bunking significantly reduced belief in election rumours (effects persisting, though attenuated, one week later) and modestly offset declines in confidence in national election administration, with no evidence of partisan backlash. After finalizing the prompt, articles written with human feedback were no more effective than articles using only AI, indicating that per-rumour human effort can be substantially reduced once the prompt is in place. Within the election-integrity domain, effective misinformation inoculation can therefore be achieved at machine speed without proportional human effort, extending the workflow to other domains (e.g. public health and climate) is a natural direction for future work. To facilitate real-world use, we release an interactive demonstration that automatically drafts pre-bunking articles from trusted facts at electionbot.chat/article.