Adoption is rising across countries, while surveys cited by Technology Review show growing nervousness and opposition.
People are using AI more even as public opinion becomes more negative. The apparent contradiction makes more sense when “AI” means two different things: a tool people choose for a task, and an industry whose plans can affect workplaces, communities, and public life.
Technology Review reports that half of U.S. adults now say they use a chatbot, more than twice the number reported in 2023. One in four people says they use one every day, according to Pew figures cited by the publication.
The pattern is global. More than one-third of adults across all 38 countries in the Organisation for Economic Co-operation and Development reported using generative AI tools in the previous three months, Technology Review says. Generative AI produces content such as text or images in response to a request.
At the same time, concern is growing. Technology Review cites Pew research showing that more U.S. adults expect AI to have a negative effect on them personally and on society than expect a positive effect. The pessimism is strongest among younger people, according to the commentary.
More than half of people worldwide say AI products and services make them nervous, Technology Review reports, citing a Stanford University report. In a May Gallup poll, 71% of U.S. adults said they would oppose a new AI data center in their area. That compared with 53% who said they would oppose a new nuclear power plant.
These findings do not measure exactly the same people or questions. The user counts, survey results, and definitions come from different sources and populations. They show that use and concern can coexist, but they do not establish how many people both use AI and oppose it.
Useful tools, unwanted pressure
Technology Review’s commentary offers one main explanation: people may object less to the underlying technology than to the companies’ relentless push to expand it.
The commentary argues that AI companies tell the public to prepare for major social and economic upheaval. That message can make people uneasy, even if they still find particular AI tools worth using.
This helps separate personal behavior from public judgment. Someone might use a chatbot while disliking the way companies promote AI or resisting a large data center nearby. Using a product does not automatically mean supporting the business strategies behind it.
The same person can also hold mixed views about the technology itself. A tool may be convenient in one situation and unwelcome in another. Public opinion does not need to settle into a simple “for” or “against” position.
The Springboards chief executive, whose startup is developing a large language model, described his company to Technology Review as “a self-loathing AI company.” He added, “We don’t know if we really like what we’re doing.”
A large language model, or LLM, is an AI system trained to generate and work with language. The executive’s comments capture the tension: companies can believe these systems are becoming unavoidable while still questioning what they should do with them.
Technology Review reports that ChatGPT reached one billion monthly users in May, according to market analysis firm Sensor Tower. Google DeepMind’s Gemini was logging 950 million users in July, according to the same commentary.
Those numbers indicate broad reach, not approval. A monthly user might use a service frequently, occasionally, or for reasons that say little about their opinion of the company. The figures also come from a market-analysis estimate, not the same survey that measured public nervousness.
Why adoption can continue
People often judge a tool by the immediate task in front of them, while judging an industry by its wider effects. Those judgments can move in opposite directions.
A chatbot may seem worth using for a specific need, while the direction of AI development may feel imposed from above. A person can accept limited, voluntary use but reject the idea that every workplace, service, or community should adopt the technology.
Technology Review compares this pattern with earlier technology battles involving social media and Google search. Those services became deeply embedded in daily life even as criticism of the companies behind them grew.
AI may be following part of that path, but the comparison does not prove that its future will be the same. The available figures show expanding use and negative sentiment; they do not explain every reason people hold those views.
The commentary also points to a difference from earlier platforms: people may still have more influence over AI’s direction. Technology Review reports that all 50 U.S. states have either passed or proposed laws governing AI development and deployment. Together, those efforts create a patchwork of more than 2,100 bills, the publication says.
The commentary also notes the presence of open-source alternatives to products from Google, OpenAI, and Anthropic. Open-source software makes important parts of a system available for others to inspect, change, or build on, although the specific freedoms depend on the project.
More choices could give users and businesses greater leverage. Regulation could also shape where and how AI systems are deployed. Neither result is guaranteed, and the supplied figures do not show whether the state laws have changed AI development or use.
The practical answer
People are using AI because adoption can offer enough value, access, or convenience to continue, even when enthusiasm is weak. They may be judging a specific tool separately from the companies’ broader plans and public claims.
That is why rising use does not necessarily signal rising trust. The next useful question is not whether people “like AI” in general, but which uses they accept, which demands they reject, and whether companies and lawmakers respond to those boundaries.
As the Springboards chief executive told Technology Review, there is “no walking back from LLMs—but you could still make them do something different.” For users, the important issue is whether that difference becomes visible in the products and rules they encounter.