What People Actually Fear — and Hope — About AI, and What the Evidence Shows

A public that is wary, but not uniform

When researchers ask ordinary people how they feel about artificial intelligence, the most consistent finding is ambivalence. In the United States, Pew Research Center found that 50% of adults felt more concerned than excited about the growing use of AI in daily life, while only 10% felt more excited than concerned and 38% felt both equally (June 2025). Concern has climbed steadily — from 37% in 2021 — as the technology has become more visible.

Globally, the mood is more balanced. The Ipsos AI Monitor, which tracks attitudes across roughly 30 countries, found excitement and nervousness at near parity: about 52% of people said they were excited about products and services using AI, against 53% who said AI made them nervous. Notably, these are often the same people. The survey frames the reactions as overlapping rather than opposing — many individuals hold both at once.

Where the worry concentrates

Several fears recur across studies. Jobs rank near the top: in Pew's work, 64% of U.S. adults expected AI to lead to fewer jobs over the next 20 years, and only 23% thought AI would improve how people do their work.

Misinformation is a second cluster. Ahead of the 2024 U.S. election, 57% of Americans said they were highly concerned that AI would be used to create and spread false or misleading information about candidates — a worry shared almost equally by Republicans (56%) and Democrats (58%). More broadly, 39% expected AI to be used mostly for bad purposes during the campaign, versus 5% who expected mostly good.

A third strand is subtler: the erosion of human capacities and connection. In a 2025 Pew study, roughly half of Americans worried AI would worsen people's ability to think creatively, and a similar share feared it would damage meaningful relationships. Privacy and a general sense of lost control also feature — more than half of U.S. adults, and a majority of AI experts, said they wanted more control over how AI is used in their lives.

Where the hope lives

Optimism is real but uneven, and it tends to attach to concrete, visible benefits. Health and medicine are the clearest example: Americans were more likely to expect AI to have a positive (44%) than negative (19%) effect on medical care over the next 20 years. People also showed openness to AI for data-heavy tasks such as weather forecasting and scientific analysis.

Practical time-savings register too. In the Ipsos data, 62% of workers said AI had saved them time in the past year, with higher earners (70%) and younger workers reporting the largest gains. Broadly, enthusiasm tracks perceived economic benefit: the countries most excited about AI — concentrated in Asia and Latin America — are generally those where people believe it will help their economies, while Europe and North America skew more cautious.

The trust gap

Underlying much of this is a durable trust deficit. A December 2025 YouGov poll found fewer than one in five Americans (18%) would trust an AI system to make a decision or take an action even somewhat, against 53% who would not. A Gallup–Bentley University survey found 77% of Americans distrust both businesses and government to use AI responsibly. Trust is also conditional: the Reuters Institute reported that people are far more skeptical of AI-generated answers on high-stakes topics like health and politics, where many still verify through traditional sources — even as everyday use has surged, with those who have ever used a generative AI tool rising from 40% to 61% in a single year.

Experts, meanwhile, are consistently more optimistic than the public. In Pew's paired survey, 56% of AI experts expected a positive impact on the U.S. over 20 years, versus 17% of the general public — a gap worth keeping in view, since expert enthusiasm does not automatically transfer to those living with the technology.

Reading the evidence honestly

These numbers describe attitudes, not outcomes, and they carry real limits. Question wording shapes answers; a vague term like "AI" can mean chatbots to one respondent and autonomous weapons to another. Sentiment shifts quickly as tools and headlines change, so figures a year apart may not be strictly comparable. And most large surveys are strongest in wealthier countries, leaving much of the world's experience thinly measured.

What the evidence does support is modest and steady: the public is neither uniformly fearful nor uniformly hopeful. People tend to welcome AI where they can see a tangible payoff — a faster diagnosis, an hour saved — and grow wary where the stakes are high, the mechanisms are opaque, or control feels out of reach. The clearest signal across all this data is not a verdict on AI, but a request: most people want more say in how it is used.