Using AI to Thwart Election Misinformation

Caltech News reports that a team of Caltech researchers has built an artificial intelligence system that automatically crafts “pre‑bunking” messages—short articles that warn voters about false election claims before they encounter them. In a pre‑registered study published in Royal Society Open Science, the AI‑generated pre‑bunks performed as well as human‑crafted ones, reducing belief in five common myths about the twenty twenty four election among a sample of more than four thousand registered voters and preserving confidence a week later. The approach scales because the model uses a reusable expert prompt and verified election data, allowing rapid creation of new pre‑bunks without additional human review. The researchers have released a public demo at electionbot dot chat and aim to partner with election officials so the tool can be deployed as rumors emerge ahead of the upcoming twenty twenty six midterms.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton