Artificial intelligence has a popularity problem. People increasingly say they distrust it, fear its impact on society and resent the companies pushing it into everyday life. At the same time, AI adoption is accelerating at extraordinary speed.
That contradiction may be one of the defining features of the current AI boom. Public enthusiasm for the technology is cooling just as tools such as ChatGPT and Gemini are becoming part of people’s daily routines.
AI Anxiety Is Rising
The numbers paint an increasingly negative picture of public sentiment. Pew Research Center polling shows more Americans expect AI to have a negative rather than positive effect on them personally and on society, with younger adults particularly pessimistic.
More than half of people worldwide say AI products make them nervous, according to research cited by MIT Technology Review. A Gallup poll found an even more striking example of the backlash:71% of US adults said they would oppose construction of an AI data center in their area, compared with 53% who would oppose a nuclear power plant.
Yet none of this appears to be stopping adoption.
Sensor Tower estimates that ChatGPT reached one billion monthly users in May, while Google’s Gemini recorded around 950 million users in July. Pew research also indicates that half of American adults now use chatbots, more than twice the proportion reported in 2023, with one in four saying they use them every day.
Across the 38 OECD countries, more than a third of adults report having recently used generative AI.
Maybe People Don’t Hate AI — They Hate the AI Industry
One explanation is that the public is distinguishing between the usefulness of the technology and its broader consequences.
Someone can use an AI assistant to summarize documents, write code or research a purchase while simultaneously worrying about job displacement, misinformation, surveillance, energy consumption or the concentration of power among a handful of technology companies.
MIT Technology Review argues that much of the frustration may therefore be directed less at AI itself than at the relentless way the industry is promoting it.
AI features are appearing in search engines, phones, operating systems, productivity software and social networks, often whether users actively requested them or not. Meanwhile, technology executives routinely describe AI as a transformation capable of reshaping entire professions and economies.
That combination creates an unusual message for consumers:this technology may radically disrupt your life, but you should also start using it immediately.
It is hardly surprising that people can find AI useful while resenting the way it is being introduced.
We’ve Seen This Relationship With Technology Before
Social media followed a similar trajectory.
Facebook, Twitter and other platforms attracted billions of users while simultaneously becoming targets of growing criticism over privacy, misinformation, mental health and political influence. People could dislike the companies behind those platforms while continuing to use the products because their friends, communities and professional networks were already there.
AI may be developing the same complicated relationship with its users.
There is an important difference, however. Generative AI remains a relatively young market, and users still have meaningful alternatives. Open-source models compete with commercial systems from OpenAI, Google and Anthropic, while new companies continue experimenting with different approaches to how AI products should behave.
Regulators are also moving considerably earlier than they did during the rise of social media. All 50 US states have either introduced or passed legislation concerning AI, producing more than 2,100 bills, according to figures cited by MIT Technology Review.
AI’s Biggest Problem May Be Trust, Not Adoption
The technology industry does not currently appear to have difficulty convincing people to try AI. The harder challenge is convincing them that the future being built around it is one they actually want.
That distinction matters.
A billion people using ChatGPT does not necessarily mean a billion people support the economic, political and social vision surrounding generative AI. Usage measures utility; it does not automatically measure trust.
The next stage of the AI competition may therefore involve more than producing smarter models. Companies will need to demonstrate that their products respect user choice, communicate limitations clearly and provide benefits without demanding unlimited trust in the organizations behind them.
The apparent contradiction between rising AI adoption and falling public enthusiasm may not be contradictory at all.People can dislike where the AI industry is heading while still finding the technology too useful to ignore.