40% of US Adults Expect AI to Harm Society, While 24% Use Chatbots Daily
In a new US survey, more people said AI would be bad for society than said it would be good. Meanwhile, chatbot use kept climbing.
Public concern about AI and everyday use of AI are emerging side by side. MIT Technology Review tackled the phenomenon in an October 5 column. In a Pew Research Center survey, 40% of US adults took a negative view of AI’s long-term impact on society. At the same time, roughly half said they use AI chatbots.
How Americans See AI’s Effect on Society
When Pew Research Center surveyed US adults in February 2026, 16% said AI would have a good effect on society over the next 20 years. The share saying the effect would be bad was 40%, 24 percentage points higher. Asked about society as a whole, in other words, concern outweighed optimism.
The answers shifted somewhat when respondents were asked about the impact on their own lives. There, 31% expected a bad effect and 23% expected a good one. The gap between the two outlooks narrowed when the question turned personal. Among those aged 18 to 29, negative expectations ran higher than the overall average for both society and personal life.
One in Four Uses a Chatbot Every Day
Chatbots have already become part of daily life for a sizable share of US adults. In the same survey, 24% said they use an AI chatbot every day. Even amid widespread worry about long-term effects, close to one in four people is opening a chatbot daily.
Source: Google (press release)
The share who said they had tried ChatGPT rose from 18% in 2023 to 44% in 2026. The two surveys, however, phrased the question differently. In the 2026 survey, ChatGPT was the most widely tried chatbot at 44%, followed by Gemini at 24% and Copilot at 17%.
If People Are Uneasy, Why Do They Keep Using It?
Other surveys likewise show feelings about AI pulling in more than one direction. In Stanford University’s 2026 AI Index report, 52% of respondents worldwide said AI products and services make them nervous. At the same time, the share saying the benefits outweigh the drawbacks rose from 55% in 2024 to 59% in 2025. Unease and an acknowledgment of the benefits showed up together.
Concerns about AI have also spread to the neighborhood level. In a Gallup survey in the United States, 71% of respondents opposed building an AI data center in their own area. Whether someone uses a chatbot, what effect AI will have on society, and whether to accept related facilities nearby are being treated as separate questions.
The MIT Technology Review columnist linked people’s continued use of AI, despite their dislike of it, to the way companies have rolled the technology out. The argument is that it matters not only how often people use AI, but also how companies push it into everyday life.
In the end, this survey found that more people are using chatbots, yet a large share still expect AI to harm society. The question worth watching from here is which benefits keep users coming back, and which changes leave them uneasy.