Gallup Finds Positive AI Feelings Lead in 34 of 37 Countries, but Not in the U.S.

AI’s global reception looks less divided than the U.S. debate might suggest. In new Gallup–Microsoft research covering 37 countries and territories, positive feelings outweighed negative ones almost everywhere. The United States was one of three exceptions, and 74% of U.S. adults aware of AI said it made them feel worried. That finding is about emotion, not a vote to stop using the technology.

Three exceptions to a positive balance

The other exceptions were Egypt and the State of Palestine. Elsewhere, a positive emotional response prevailed, but that does not mean worry was absent. Gallup asked people aware of AI whether it made them feel curious, happy, excited, worried, sad or angry. Respondents could report more than one feeling, so curiosity and worry are not opposing camps.

Worry was particularly common in wealthy Western countries. Seven Western high-income countries surveyed occupied six of the top seven places for worry and the eighth place. The U.S. led at 74% of AI-aware adults, followed by the Netherlands at 67% and Canada at 64%. Yet these countries were also relatively heavy users: a median of 29% used AI daily, compared with 17% across the other 30 countries.

45%Daily AI users

In the U.S., 45% of daily AI users said they trusted AI completely or a lot to provide accurate information.

8%Never used AI

Among U.S. adults who had never used AI, 8% expressed that level of trust in its accuracy.

Use and trust move together—not necessarily in sequence

The trust gap extends through the U.S. usage groups: 26% of weekly users and 15% of less-than-weekly users trusted AI’s accuracy completely or a lot. Worry follows a different measure but a related pattern. It affected 68% of U.S. daily users, 80% of those using AI less than weekly and 74% of people who had never used it. Daily use did not erase concern, even among the group most inclined to trust AI’s answers.

Those comparisons cannot settle whether experience builds trust, whether people who already trust AI choose to use it more, or whether both happen. Gallup makes that limit explicit. Trust also has a narrow meaning here: the question concerned accurate information, not whether AI is safe, fair or protective of privacy. Treating the results as a general endorsement of AI would stretch them beyond what was asked.

Hope can outrun firsthand use

Feeling good about AI, expecting it to help and trusting its answers were separate questions. In 32 countries, more AI-aware adults expected the technology to improve their own daily lives than to worsen them. In 35 countries, more expected AI to help their country than trusted it to provide accurate information. Across the surveyed countries, the median share expressing a lot or complete trust in its accuracy was 36%.

Nor did favorable expectations require extensive firsthand use. In Nigeria, 68% of people aware of AI expected it to mostly help their country, although 20% of adults had ever used it. Vietnam showed an even wider split: 91% expected help, while 43% had ever used AI. The U.S. sat near the bottom on that expectation, at 36%. These are country results, not evidence that use caused any country’s outlook.

The survey is still growing

These are the first 37 countries and territories to complete fieldwork in a Gallup–Microsoft project intended to cover 140. Gallup surveyed nationally representative samples of roughly 1,000 people ages 15 and older in each place from April through July 2026. Across the countries surveyed so far, the median share aware of AI was 81%, but the median share who had ever used it was 43%. That leaves a large group forming opinions without firsthand experience.

The remaining countries may change the picture of global opinion. For now, the sharpest contrast lies within this first group: widespread positive feelings alongside unusually high worry in countries where AI use is already relatively common.