Download the Full Report: Public Perceptions of AI - August 2025

AI’s Trust Deficit and Pathways Toward Alignment

National polling by Civiqs conducted from April 2025 to August 2025 finds that Artificial Intelligence has a broadly held public trust problem. Americans express a range of concerns towards AI technology that go beyond any single company or application.

Most people do not trust AI to be safe or helpful, and, as AI becomes more widespread, Americans increasingly believe that Artificial Intelligence is having a negative impact on the world.

Tracking attitudes towards AI is critical to ensuring alignment—research that Civiqs specializes in providing. The findings in this report offer a benchmark for bringing AI into alignment with social preferences, addressing public skepticism, and guiding AI development in ways that people find valuable and trustworthy.

Key Findings:

• Just over half of American adults have used an AI chatbot; mainly for personal rather than professional purposes. Adoption is highest among younger users.
• Sentiment towards AI tends to be unfavorable: More Americans believe that AI is generally a bad rather than good thing (44% bad, 35% good), and that overall, AI is having a negative impact on the world (49% negative, 18% positive).
• People who use AI most frequently—including on a daily basis—have highly positive opinions of the technology, but this is a relatively small group.
• Two thirds of Americans do not trust AI to be safe, and half do not trust it to be helpful.
• In their own words, the primary reasons Americans say AI is unsafe is because it produces inaccurate information, is susceptible to malicious use by bad actors, and is not subject to sufficient control and regulation.