Democrats and Republicans back AI job retraining in Northeastern survey

Support for learning to work with AI is not the same as confidence in the industry behind it. A Northeastern University survey report released October 23 found that 64% of Republican respondents and 63% of Democrats backed AI-related job retraining, even as distrust of AI companies cut across political groups.

The survey covered more than 38,000 people across all 50 states between June and August 2026, with respondents from different demographic groups and political parties. It was conducted through the Civic Health and Institutions Project, known as CHIP50, a multiversity research initiative. The findings therefore describe views collected over the summer, rather than a snapshot taken at the report’s October release.

Agreement on training—and keeping it portable

The retraining question concerned helping workers gain the skills needed for jobs that make heavy use of AI. Researchers said that could include both blue-collar and white-collar work. Support was nearly identical among Republican and Democratic respondents, making workforce preparation a point of agreement despite their different levels of trust in AI companies.

A separate proposal also drew majority support in both parties: tax-advantaged savings accounts that workers could use for training and education. These “worker-owned retraining accounts” would stay with people when they changed jobs. Fifty-seven percent of Republicans and 56% of Democrats supported the idea. That question went beyond whether training was desirable to ask about a specific way workers could pay for it.

David Lazer, a Northeastern professor of political science and computer sciences and one of the lead researchers, described the agreement as the survey’s most surprising result. His interpretation was that partisan differences over trust had not prevented agreement on a government role in protecting and preparing workers for AI-related change.

We are seeing an increasing partisan divide in terms of trust, but some bipartisan consensus about the need for the government to protect people from the incoming storm that AI will unleash.

David Lazer, Northeastern professor and survey lead researcher

Using AI without trusting its makers

The trust findings were less bipartisan than the training results. Independents were the most distrustful political group, while Republicans were evenly divided. The question concerned AI companies—not whether respondents trusted every answer from an AI tool or believed every use of the technology was harmful.

• Independents: 66% said they distrusted AI companies, the largest share among the political groups.
• Democrats: 61% said they distrusted AI companies, also a majority.
• Republicans: 50% trusted AI companies and 50% distrusted them.

Young adults showed a particularly clear gap between experience and confidence. Among respondents aged 18 to 24, 89% said they had used AI tools, but only 30% said they trusted AI companies. Those figures distinguish trying the technology from endorsing its makers. They do not, by themselves, explain why an individual respondent used AI despite having doubts about the companies.

A majority rejects a local data center

On infrastructure, the survey asked about an AI data center in respondents’ own community. Fifty-four percent said they would not support one, compared with 22% who would; 23% had no opinion. Opposition was therefore more common than either support or uncertainty. This was a measure of willingness to host a facility locally, not a vote on a particular construction project.