Comparing Trust: Government AI vs. Private-Tech, Foreign Democracy, Party-in-Power

In isolation, the question of whether Americans would trust a government AI – a system engineered and operated entirely by the U.S. government – yields a relatively even split between acceptance and rejection. The only problem with that question?

Trust does not exist in isolation.

Ask the same people to choose between government and private-sector or foreign-democracy AI, or to imagine it under different political conditions. Suddenly, the numbers shift, painting an entirely different, and often surprising, picture.

How does trust in government AI change with the context?

Sometimes dramatically. Sometimes subtly. Sometimes not at all. The public trust in government AI is not a stable number – it shifts depending on what the alternatives are and which context is applied. 

In our May 2026 survey, we tested three forms of contextual pressure:

• Sector: Who built the AI – private tech or the government?
• Origin: Where does the AI come from – a foreign democracy or the U.S.?
• Party-in-power: Which political party currently controls the federal government?

Each pressure factor type produced measurable change in trust – and in some cases, a measurable lack of change.

Would Americans trust government AI more than private-tech AI?

As a matter of fact, no. Only 12% of respondents stated they would trust a government AI more than a private-tech one, while ~31% said they would trust it less. The gap leans toward distrust nearly 2.5 times stronger. 

The directional preference remains consistent when the distribution is observed by party affiliation: 

Noticeably, across both major groups in our dataset, trust in government AI is at a net negative (by 17 and 19 points, respectively), suggesting that the directional preference for private-sector AI is structural, rather than partisan.

Interestingly, the largest segments across all three cohorts (46% general; 47% and 44% by party affiliation) don’t differentiate between the two, yet remain approximately the same - suggesting that, for nearly half of Americans, who built the system isn’t a meaningful trust differentiator on its own.

The ITIF 2026 Report shows a similar pattern. Across six government AI scenarios tested, from mass surveillance of U.S. citizens to autonomous weapons, the “intense concern” segments held within a narrow 53%-56% band, with minimal variation by use case and consistent across both party affiliations. Combined, these findings suggest the trust deficit is institutional, not political or situational.

Would Americans trust a foreign democracy’s AI more than their own government’s?

Once again, the answer is no. Even though the question explicitly specified a “foreign democracy with a strong rule of law,” more than 4 out of 10 (43%) of respondents were ready to outright reject their AI. In contrast, only 11% would trust an AI built by a foreign government more, saying that “some governments are simply more trustworthy than mine [U.S.].” 

Interestingly, though, nearly 14% of respondents are open to the idea of foreign-democracy AI, but only if it was built jointly with international oversight, indicating the presence of conditional openness - trust is available, but only if it comes with the accountability structure attached.

By party affiliation, the directional preference diverges more strongly than in the government-vs-private-tech comparison:

The Republican cohort’s sentiment is stronger across both extremes - they are more likely to outright reject (50%) or accept (17%) the foreign AI. The Democratic distribution is more centered, with a larger share (21%) saying country of origin isn’t a major trust factor, and a lower intensity on both extremes (37% reject, 9% accept).This finding is quite surprising, considering that the U.S. reported the absolute lowest trust (31%) in its own government to regulate AI responsibly, according to Pew Research data cited in Stanford HAI’s 2026 AI Index.

Does trust in government AI shift depending on which party is in power?

It does, it doesn’t, and it doesn’t matter. Here we have a notably balanced distribution: for 27% the trust would shift (somewhat or a lot), for 32% the trust wouldn’t shift (much or at all), and 33% wouldn’t trust government AI regardless of which party is in power, reflecting a broader structural distrust pattern documented across ours and external surveys.

The by-party-affiliation split is even more symmetrical, with the distribution landing within three points across every category for both major cohorts:

This is arguably the most interesting finding in this dataset: An effectively identical proportion of respondents whose trust shifts with power, those whose trust cannot be swayed, and those who reject government AI categorically - across both parties. The nearly perfect three-way split - in the area where the strongest asymmetry could’ve been expected.

Where does government AI sit in the trust hierarchy?

Externally - below private-sector AI and above foreign-government AI. Internally - at the orthocenter between conditional trust, stable trust, and categorical rejection. That’s the picture when we look at the three contextual pressure factors separately (sector, origin, party-power). When we look at all three together, however, a different image emerges:

1. Sector: In the direct head-to-head, both parties’ affiliates directionally prefer private-sector AI more; only 12% of respondents would trust government AI more.
2. Origin: In the comparison with foreign-government AI, the native AI wins by elimination, not by merit. 11% would trust foreign AI more, but the 43% who outright reject it do not redirect that trust toward the domestic AI option.
3. Party-power: The trust differential between Democrats and Republicans is the smallest measured in the survey, implying that the trust deficit in government AI is not political, but institutional.

Because the American public’s skepticism is aimed at the institution rather than leadership, the practical implications are uncomfortable for both parties:

Jordan Parkes, the CEO and founder of ZeroClick Labs, has been building and scaling digital marketing strategies since 2012, leveraging performance-driven SEO and data-based digital marketing solutions to guide the growth of hundreds of companies across the U.S. and Europe.