Americans’ attitudes toward artificial intelligence are souring, and the people who know the most about the technology are often the most uneasy about it, according to new data from Gallup.
After two years in which the public had been gradually warming to AI, Gallup now finds that the trend has reversed. The polling firm reports that about seven in ten U.S. adults say they are either “somewhat” or “extremely” knowledgeable about AI-up from 64% in 2024. But instead of translating into enthusiasm, that growing familiarity has gone hand in hand with deeper doubts about the technology’s benefits and how it is being used.
A rising share of Americans now believes AI is a net negative. Thirty‑nine percent say artificial intelligence does more harm than good, compared with 31% in 2025. Only 9% say AI brings more good than harm. The majority, 52%, see it as a mixed bag, doing roughly equal amounts of harm and good.
Gallup’s data suggests that, as people learn more about how AI works and where it’s being deployed, they are less inclined to give it the benefit of the doubt. The shift in sentiment is especially pronounced among those who describe themselves as highly knowledgeable about AI. For this group, awareness of real‑world use cases, from automated hiring tools to generative models that can mimic human speech and images, appears to translate into heightened concern rather than excitement.
One of the strongest drivers of this skepticism is anxiety over employment. More Americans now expect AI to reduce the number of U.S. jobs in the coming decade. The idea that AI will automate not just repetitive physical labor but also white‑collar and creative work has moved from abstract theory to something people feel in their own workplaces-through chatbots, automated customer service, algorithmic decision‑making, and productivity tools embedded in everyday software.
Alongside fears of job loss, trust in corporate use of AI is eroding. Americans are increasingly reluctant to assume that businesses will deploy AI responsibly. Concerns range from data privacy and surveillance to biased algorithms, opaque decision‑making, deepfakes, and the use of AI to manipulate consumer behavior or undermine workers’ bargaining power. The survey indicates that, for many, AI is no longer seen as a neutral tool but as a force shaped by corporate incentives that may not align with the public interest.
This skepticism does not necessarily mean that Americans reject AI outright. The majority still see both pros and cons, and many people use AI‑powered tools daily, often without thinking about them: recommendation algorithms, navigation apps, fraud detection, translation, and more. But the balance of public opinion is shifting away from optimism toward caution, especially when the technology touches sensitive areas like employment, personal data, and public safety.
The emerging pattern raises a crucial issue for policymakers and industry leaders: as AI becomes more powerful and more embedded in critical systems, public consent is not guaranteed. People want to know who is accountable when AI gets things wrong, how their data is being used to train these models, and what guardrails exist to prevent abuse. Without clear answers and strong protections, rising familiarity with AI may continue to fuel distrust.
For businesses, Gallup’s findings are a warning sign. Companies rushing to adopt AI for cost savings or efficiency gains risk a backlash if employees and customers perceive that the technology is being used primarily to cut jobs, harvest data, or shift responsibility away from humans. Transparent communication about how AI is used, independent audits for bias and safety, and visible investments in worker retraining and upskilling can make a difference in how these tools are received.
Education also plays a double role. On one hand, teaching people how AI works-its limitations, failure modes, and appropriate use cases-can help dispel myths and prevent both fear‑mongering and overhype. On the other, as Gallup’s numbers suggest, better understanding can reveal legitimate risks that justify a more critical stance. A more informed public may be less dazzled by marketing promises and more insistent on accountability, regulation, and ethical standards.
Regulators are already trying to catch up. Discussions about AI rules increasingly focus on transparency (making it clear when AI is being used), data protection, nondiscrimination, and safety testing before deployment. Public opinion trends like those identified by Gallup are likely to strengthen political pressure for stricter oversight, especially around workplaces, schools, healthcare, financial services, and law enforcement.
For individuals, the survey underscores the importance of engaging with AI not as a distant, abstract technology but as something that will shape careers, economic security, and civic life. That means asking practical questions: How might AI change my job and my industry? What skills will be more valuable in an AI‑saturated economy? How can I use AI tools to augment my work rather than be replaced by them? And what kinds of rules and norms do I want my elected representatives to support?
Ultimately, Gallup’s findings do not point to a simple story of “pro‑AI” versus “anti‑AI.” Instead, they reveal a public that is growing more sophisticated in its view: more aware of AI’s real capabilities, more attuned to its downsides, and less willing to accept its rapid spread without conditions. As knowledge of AI expands, enthusiasm is giving way to a more skeptical, demanding, and nuanced public conversation-one that governments and companies will increasingly have to address if they want their AI initiatives to be seen as legitimate.