Gallup finds daily AI use does not ease US anxiety

A 37-country survey

Gallup surveyed roughly 1,000 people in each of 37 countries between April and July. Microsoft commissioned the research, which is planned to expand to 140 countries. India, Australia and Malaysia are not included in the published results.

The US sits among the more anxious Western countries. Across all Americans familiar with AI, 74% reported concern, while only 36% said the technology would mostly help their country.

Canada shows a similar combination of adoption and unease:

That concern is not purely abstract. Technology companies have laid off hundreds of thousands of employees to invest in AI, while entry-level vacancies have declined. There have also been reports of AI-linked psychosis, and company leaders have warned that large numbers of jobs could disappear. Anthropic CEO Dario Amodei was among them.

The trust gap

The survey’s more positive results come from countries where AI use is more widespread. Singapore had the highest share of daily users, at 46%, and more than 80% of people familiar with AI there said it would improve their daily lives. Another 77% expected it to benefit the country.

In China, 35% reported using AI every day, and most respondents expected it to help the country. Israel showed a similar pattern.

Confidence in AI output is a separate matter from optimism about its effects. Daily users generally trust AI results more than infrequent users, especially in Singapore, Israel and China. In the US, only 45% of daily users said they had that level of confidence. Across all countries, the median share expressing high trust was 36%.

Anxiety is becoming part of the policy debate

Positive emotions toward AI are more common overall. Gallup says curiosity was the most common response in 34 of the 37 countries, while about 32% of all respondents reported anxiety. Positive feelings were especially common in:

The survey does not show how attitudes may have shifted in some countries, particularly China, since fieldwork ended in July. AI safety is expected to be a major topic at Thursday’s meeting between US President Donald Trump and Chinese Premier Li Qiang. Chen Yixin, head of China’s Ministry of State Security, recently wrote that AI would have a systemic effect on the political and security environment and could threaten China’s political, institutional and ideological security.

Gallup senior research fellow Pablo Diego-Rosell told TechCrunch that attitudes toward AI have several dimensions. People can be curious about it, consider it useful and use it frequently while also feeling anxious. They can see substantial potential in the technology while doubting the accuracy of its answers.

My reading is that this is the survey’s most useful finding: familiarity does not resolve the argument over AI. It separates the argument into practical questions about where the technology should be allowed and what safeguards should apply.

People who use AI in daily life may still support stricter rules or oppose its use in hiring, healthcare, education and government. Other disputes concern the climate cost of training and running models, cybersecurity, content owners’ rights, creative industries and data privacy.

The announcement is quiet about the question that would make these concerns actionable: which specific uses do daily users want restricted, and which do they consider acceptable? As investment keeps pushing AI into more parts of society, broad adoption may grow alongside a public that is using the technology without granting it broad permission.