Election misinformation raises new concerns as AI chatbots give voters wrong answers
Experts warn that false voting claims, immigration fears and language barriers could discourage eligible voters ahead of the November 3 election.

Experts warn that false voting claims, immigration fears and language barriers could discourage eligible voters ahead of the November 3 election.

As voters prepare for the Nov. 3 midterm elections, misinformation circulating through social media and artificial intelligence platforms is raising concerns about whether Americans, particularly those in immigrant communities, are receiving accurate information about their voting rights.
During an Oct. 23 national news briefing hosted by American Community Media, researchers warned that false claims about election fraud, immigration enforcement and voter eligibility are spreading across digital platforms, sometimes reaching audiences in their native languages without adequate fact-checking.
The briefing, Who Gets the Last Word? Election Disinformation in Immigrant Communities, brought together experts examining how misinformation spreads and why language barriers can make some communities particularly vulnerable.
One of the most troubling findings involved AI chatbots, which a growing number of people use to answer everyday questions.
Laura Zommer, co-founder and CEO of Factchequeado, cited research by the Institute for Strategic Dialogue examining how major AI chatbots respond to questions about U.S. elections.
According to the research she presented, approximately 30% of English-language responses were incorrect, incomplete or outdated. In Spanish, that figure climbed to 45%.
In one example, Zommer said Meta’s chatbot twice provided Nov. 4 as the date of the 2026 midterm elections when asked in Spanish. The correct date is Nov. 3.
The errors involved basic election information, not complicated political questions.
Zommer also noted that AI chatbots cited fewer sources in Spanish than in English, raising questions about the quality and reliability of information available to Spanish-speaking voters.
Texas Under the Microscope
The findings carry particular significance for Texas, where voters from diverse linguistic and cultural backgrounds are participating in the election.
Zommer announced that Factchequeado is conducting additional research specifically focused on Texas, examining how AI chatbots respond to election-related questions in English and Spanish.
The project also explores whether responses change depending on how political questions are framed.
Preliminary findings, she said, suggest that chatbots may provide answers that align with what users appear to want to hear, potentially reinforcing existing beliefs rather than correcting inaccurate information.
The Texas-focused report was expected to be released Oct. 30.
Immigration Fears and False Voting Claims
Cameron Hickey, CEO of the National Conference on Citizenship, warned that election misinformation extends beyond misleading political advertisements.
He described AI-generated images depicting immigration enforcement agents intimidating voters at polling places. Although the scenes were fabricated, Hickey said such images were circulating online and could heighten existing fears among immigrant communities.
Hickey also expressed concern about reports of actual immigration enforcement activity near polling locations, saying the combination of real events and fabricated imagery could discourage participation.
He raised additional concerns about claims involving noncitizen voting and citizenship verification systems, noting that inaccurate or outdated records can create confusion about the eligibility of naturalized citizens.
Jinxia Niu, who leads the Chinese Digital Engagement Project at Chinese for Affirmative Action, said immigration-related misinformation can have consequences even for people legally entitled to vote.
“It really turns this disinformation into immigration fear,” Niu said, adding that her organization’s work suggests such fears are already discouraging some eligible citizens from participating.
Niu also described a shift in Chinese-language election misinformation from WeChat toward X, formerly Twitter. She said misleading claims originating on X are often translated and shared through WeChat groups and other platforms.
Chinese for Affirmative Action has identified more than 150 social media accounts associated with disinformation in the Chinese-language information environment, according to Niu.
A new report, Divided by Design, examines how anti-immigrant narratives are creating divisions within Chinese, Indian, Filipino and Vietnamese communities.
What Voters Can Do
For journalists and voters trying to separate fact from fiction, Hickey recommended a four-step verification method known as SIFT, developed by information literacy researcher Mike Caulfield:
Hickey cautioned that a large social media following does not necessarily indicate credibility.
He also emphasized that misinformation is not exclusively a technological problem. False narratives can undermine confidence in elections, public institutions and even neighbors.
For Houston-area voters, the message is particularly timely as Election Day approaches.
Voters can verify registration information, polling locations and election procedures through the Texas Secretary of State and their county election offices rather than relying solely on social media posts or AI-generated answers.
As Nov. 3 approaches, the experts stressed that accurate information, accessible in the languages communities use, remains essential to protecting public confidence and participation in the democratic process.
[wtpsw_gridbox]
