Back to AI & Tech

How Americans want AI handled in politics and public life

Overview

Americans bring clear expectations and real wariness to AI's move into politics. Strong majorities want disclosure when campaigns use it, a plurality would penalize candidates who do, and most doubt they have any influence over how it is governed.


The wariness is not limited to elections. About 63% say AI's main beneficiaries are wealthy investors and big corporations rather than ordinary Americans, and 31% expect AI to shape judges' decisions to a large extent.

Topline

73.7% say campaigns should always disclose when they use AI.

If a political campaign uses artificial intelligence to write, edit, or generate campaign messages, should the campaign be required to disclose that AI was used?

  • Yes, always 73.7%
  • Yes, but only when the content includes images, video, or audio 9.9%
  • Yes, but only when the content is misleading or deceptive 8.0%
  • Not sure 4.8%
  • No, disclosure should not be required 3.5%

2026 · base n 1,000 · +/- 3.2%

Verasight MPSA Omnibus Survey #2026-050

View source data

Most want campaigns to disclose AI use

73.7% say a campaign should always disclose when it uses AI to write or generate messages. Another 9.9% want disclosure for images, video, or audio, and 8.0% when the content is misleading, so more than nine in ten favor disclosure in some form.

Only 3.5% say disclosure should not be required, with 4.8% unsure.

Stacked breakdown

44.7% would be less likely to support a candidate whose campaign used AI.

If you discovered that a political campaign used AI to write materials like campaign websites, advertising materials, and scripts for canvassers, would you be more likely, less likely or equally likely to support the candidate?

Much less likely
23.8%
Somewhat less likely
20.9%
Neither more or less likely
44.3%
Somewhat more likely
7.0%
Much more likely
4.0%

2026 · base n 1,000 · +/- 3.2%

Verasight MPSA Omnibus Survey #2026-050

View source data

Using AI carries a penalty with voters

If they learned a campaign used AI for materials like websites, ads, and canvassing scripts, 44.7% say they would be less likely to support the candidate, including 23.8% much less likely and 20.9% somewhat less likely.

Just 11.0% would be more likely to support such a candidate, while 44.3% say it would make no difference.

Stacked breakdown

78.4% feel they have little or no say in how AI is regulated.

To what extent do you feel you have a say in how artificial intelligence is developed and regulated by companies and government?

Not at all
48.7%
Not much
29.7%
Somewhat
14.4%
Very much
7.1%

2026 · base n 1,000 · +/- 3.2%

Verasight MPSA Omnibus Survey #2026-050

View source data

Wariness extends beyond the campaign

78.4% feel they have little or no say in how AI is developed and regulated, with 48.7% saying not at all and 29.7% not much.

62.7% agree the main beneficiaries of AI are wealthy investors and big corporations, not ordinary Americans.

On AI in the courts, 31.3% think it will affect judges' decisions to a large extent, while 49.8% expect a minimal effect.

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
2026-06-21 → 2026-06-21
Base (unweighted)
1,000
Margin of error
+/- 3.2%
Module
3
Sponsor
Verasight
Weight variable
weight
Weighting targets
age, race/ethnicity, sex, income, education, region, metropolitan status

Sources

[5]

Citation

Verasight MPSA Omnibus Survey #2026-050, fielded June 21-21, 2026, N=1,000 US adults age 18+, +/- 3.2%.

https://reports.verasight.io/r/mpsa26#q-128

Verasight survey methodology

How Verasight conducts surveys.

This page describes the Verasight general survey contract, separate from how the Data Library packages it. When a wave report is published, its field dates, sample sizes, and module breakdown are listed in that report.

Mode
Verasight panel recruited via random address-based sampling, random person-to-person text messaging, and dynamic online targeting.
Population
US adults age 18+.
Sample design
Surveys are run as omnibus or single-topic waves. Omnibus waves are split into modules with their own respondent set, typically around one thousand respondents per module.
Field window
Each wave specifies its own field dates. Most omnibus waves field across roughly two weeks.
Weighting
Per-module weighting to CPS targets including age, race and ethnicity, sex, income, education, region, and metropolitan status.
Partisanship benchmark
Pew Research Center's NPORS benchmarking surveys, three-year running average.
Vote benchmark
2024 presidential vote population benchmarks.
Margin of error
Typically about plus or minus 3.4 to 3.6 percent per module at standard module sizes. Question-level MoE is recomputed when a base shrinks materially below the module baseline.
Reporting
The canonical report hub is reports.verasight.io. Verified legacy reports remain on their original hosts, and unpublished reports are marked pending.
Transparency
Verasight is a member of the American Association for Public Opinion Research Transparency Initiative.

The canonical report hub is reports.verasight.io; older reports may remain on verified legacy hosts.