Qlik Survey: American Voters Trust AI Chatbots for Election Information More Than They Use Them

National survey of registered voters finds a persistent gap between where Americans get election information and where they say they'd trust it most, as AI chatbots and government sources have more trust than adoption.

Key takeaways:

• Social media is the least trusted source Americans rely on most: 48% of voters use social media for election information – more than any other source – but only 23% call it trustworthy, the widest use-trust gap of any channel measured.
• AI trust is already ahead of AI adoption: 11% of voters currently use AI chatbots like ChatGPT, Claude, Gemini, or Copilot to consume election information, but 13% name AI as a trusted source.
• Most voters trust AI, but still double-check it: 39% say they’d trust an AI explanation of a candidate or ballot measure but would verify it elsewhere; meanwhile, Gen Z is five times more likely than Baby Boomers to trust AI outright with no verification (20% vs. 4%).
  

PHILADELPHIA, Sept. 29, 2026 (GLOBE NEWSWIRE) -- Qlik^® today announced new survey findings on how registered voters are navigating an increasingly crowded and confusing information landscape ahead of the 2026 midterm elections. The research finds that Americans’ most-used source for election news, social media (Facebook, LinkedIn and Instagram, etc.), is also their least trusted. Conversely, institutional and AI-based sources are earning trust faster than they are earning usage.

Heading into the midterms, voters have more ways than ever to follow an election, from traditional news outlets to social media, podcasts, search engines, campaign content, and now AI-generated summaries. That additional access hasn’t translated into more confidence: more than a third (36%) of voters say they feel overwhelmed by the sheer volume of election information they encounter, and many are left to decipher credible information from noise largely on their own.

“Voters are getting buried in information, but not necessarily in trust,” said James Fisher, Chief Strategy Office at Qlik. “The channels people rely on most aren’t the ones they believe most. That’s the same trust gap we see in the enterprise: more data doesn’t automatically mean more confidence. It has to be earned through transparency, so people know where an answer is coming from and can trust what it tells them.”

What’s new

• Social media is the top source, but not the most trusted: Despite being used by nearly half of voters (48%) for election news, only 23% view it as a trustworthy source — a 25-point gap, the widest of any channel measured. Meanwhile, AI chatbots and government sources show the opposite pattern: roughly 1 in 10 voters (11%) currently turns to AI tools for election information, but 13% already name AI chatbots among their most trusted sources. Furthermore, only 8% turn to a government or official election website, but 16% say they'd trust it most – making AI and government sources the only two channels where trust outpaces actual use.
• Most voters take a "trust but verify" approach to AI, and fact-checking is their top reason to use it: Only 14% would take an AI's explanation of a ballot measure or candidate at face value; 39%, the largest group, say they'd trust it but still check another source. That same instinct shows up in how voters say they'd actually use AI: checking whether a candidate's claim is true or misleading is their number one use case (43%), ahead of comparing candidates' policy positions (37%) or summarizing a complex political issue (32%).
• Younger voters trust AI further, and faster: Gen Z is five times more likely than Baby Boomers to trust an AI explanation of election information without checking elsewhere (20% vs. 4%), even as Gen Z reports the highest overload from election information of any generation (43%).
• Voters trust AI with their job more than their vote: 68% of voters use AI tools regularly at work to find information or complete tasks, but only 48% have ever used AI to learn more about a political issue, candidate or election — a 20-point gap between professional and civic adoption.
• Men and women approach AI in politics very differently: Men are 16 points more likely than women to have used AI to research a candidate, issue, or election (55% vs. 39%), and women are far more likely to opt out entirely – 38% say they'd never use AI for this, compared to 24% of men. That gap carries into trust: women are also more likely to reject AI for election information outright (20% vs. 14% of men).
  

These findings point to a voter base absorbing more election information than ever while withholding trust from the channels it relies on most. Institutional sources like government websites and AI tools are earning trust ahead of adoption, suggesting real room to grow if that trust is met with transparency. Verification remains the norm for most voters, though that habit narrows sharply among younger generations, who are both the heaviest users of AI for political information and the least likely to double-check what it tells them.

Methodology

The research was conducted by Censuswide, among a sample of 2,000 respondents (18+) who are registered to vote in the United States. The data was collected between 11.09.2026 - 15.09.2026. Censuswide is a member of the Market Research Society (MRS) and the British Polling Council (BPC), and a signatory of the Global Data Quality Pledge. We adhere to the MRS Code of Conduct and ESOMAR principles

About Qlik

Qlik helps teams get more out of AI with data they can rely on and control. It delivers trusted data products, a powerful analytics engine, and AI agents. This helps teams reduce risk, keep operating costs in check, and scale AI responsibly as needs evolve. Used by 75% of the Fortune 500, Qlik supports customers worldwide. Qlik works with the systems and partners customers already use, so teams can stay flexible without lock-in.

Media Contact

Keith Parker 
512-367-2884

© 2026 QlikTech International AB. All rights reserved. All company and/or product names may be trade names, trademarks and/or registered trademarks of the respective owners with which they are associated.