Few Americans use large language models to get political information
Core finding
A majority of Americans (53%) say they never use large language models such as ChatGPT to inform themselves about politics, according to a September 2025 Verasight survey of 3,000 U.S. adults.
The rest turn to LLMs at least occasionally: 14% use them several times a week and about as many less than once a month, while 10% do so once a week and 9% one to three times a month.
Topline distribution
Few Americans use large language models to get political information.
How often do you use large language models (LLMs, ex: ChatGPT) to inform yourself about politics?
n = 1,000 · Sep 22–29, 2025 · MoE ±3.3%
Methodology
Full methodology →| Mode | Verasight panel recruited via random address-based sampling, random person-to-person text messaging, and dynamic online targeting |
|---|---|
| Population | US adults age 18+ |
| Field dates | Sep 22–29, 2025 |
| Base (unweighted) | 1,000 |
| Margin of error | ±3.3% |
| Module | Verasight APSA Omnibus Survey #2025-119 |
| Sponsor | Verasight |
| Weight variable | weight |
| Weighting targets | age, race/ethnicity, sex, income, education, region, metropolitan status |
Sources
- reports.verasight.io/r/apsa-2025
How often do you use large language models (LLMs, ex: ChatGPT) to inform yourself about politics?
Cite this question
Verasight APSA Omnibus Survey #2025-119, fielded September 22-29, 2025, N=1,000 US adults age 18+, +/- 3.3%.