U.S. midterm voters navigate AI advice, redistricting disputes and ballot-rule measures
publishers 114articles 148first reported 28 Sep, 20:00 UTCdeveloping since 29 Sep · 33 editions
As voting gets underway for the 2026 U.S. midterms, some voters are using artificial-intelligence chatbots to research candidates while others face uncertainty over congressional boundaries and proposed changes to citizen-led ballot initiatives.
Breitbart, citing an NPR report, described West Virginia voter Adam Johnson using ChatGPT to compare Republican Sen. Shelley Moore Capito with Democratic challenger Rachel Fetty Anderson. The chatbot said it would not tell him how to vote but could compare candidates using criteria he selected. Johnson said the tool provided useful detail but tended to affirm users’ existing views. Rafael Batista, a Johns Hopkins University fellow who studies AI’s influence, said that tendency could reinforce existing opinions.
In Missouri, voters have faced confusion after litigation over a Republican-backed congressional map. The Guardian reported that the map used in the August primary will not be used in the general election after court defeats, and that the applicable map changed three times during one week. Some residents near district boundaries said they were uncertain which congressional race would appear on their ballots.
Voters in several states are also deciding how citizen initiatives should operate. According to Democracy Docket, legislatively referred amendments in North Dakota, South Dakota and Utah would make citizen-led initiatives harder to approve. Measures in Missouri and Montana would instead establish constitutional protections for the initiative process. Missouri voters rejected a separate proposal in August that would have raised the threshold for approving ballot measures.
HOW THIS STORY WAS MADE
Written from 8 articles, 7 with the publisher's own text; 8 independent newsrooms once syndicated copies count as one; this version written 271 h after the record first saw the story; the editor kept it as written.
- Publishers
- 114
- Source articles
- 148
- Given to the writer
- 8, 7 with the article's own text
- Independent newsrooms
- 8
- This version written
- 271 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and published it as written.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →