Is It Rational to Be Afraid of AI? What the Surveys Show
About half of Americans say they are more worried than excited about AI, and the share has not fallen as people started using it. A look at who is afraid, what they are afraid of, and whether the fear holds up.
Is it rational to be afraid of artificial intelligence? Mostly, yes. The fears people report are specific, they are shared by about half the public, they have not faded as people started using the technology, and several of them match what the people deploying AI say they expect. What the surveys cannot tell you is how much fear is the right amount.
- 52%of Americans are more concerned than excited about AI in daily life
- 71%expect AI to lead to fewer jobs over the next 20 years
- 63%say AI is advancing too quickly
- 9%are more excited than concerned
All four figures come from Pew Research Center surveys of US adults conducted in 2026.
How many people are afraid of AI?
About half of American adults. In a Pew Research Center survey from June 2026, 52% said the growing use of AI in daily life makes them feel more concerned than excited. Another 37% said they feel both equally. Only 9% said they are more excited than concerned.
The jump came between 2022 and 2023, the first year after ChatGPT was released. Source: Pew Research Center, survey of US adults, June 2026.
Two things stand out. The concern jumped by 14 points in a single year, and then it stayed. Four years of people trying the tools for themselves have not brought the number back down.
This is not only an American mood. In Pew's 2025 survey of 25 countries, there was no country where the largest group was more excited than concerned.
Who is most afraid of AI?
Not who you might expect. The oldest adults are still the most concerned, but the youngest have caught up with almost everyone else.
| Age group | More concerned than excited, 2021 | 2026 |
|---|---|---|
| 18 to 29 | 31% | 55% |
| 30 to 49 | 34% | 51% |
| 50 to 64 | 41% | 47% |
| 65 and over | 43% | 59% |
For the first time, a majority of adults under 30 say they are more concerned than excited. Their excitement has dropped too, from 25% in 2021 to 11% in 2026.
The generation that uses AI the most is the one whose worry is rising fastest.
That matters for how the fear is read. Technophobia is usually pictured as an older person refusing to learn a new device. This is something else: people who know the tools well and are uneasy anyway.
What exactly are people afraid of?
The fears are concrete. In Pew's surveys of US adults in 2026:
- Jobs. 71% think AI will lead to fewer jobs in the United States over the next 20 years, up from 64% in 2024. Only 5% think it will lead to more.
- Privacy. 71% think the increased use of AI will make their personal information less secure. 3% think it will make it more secure.
- Speed. 63% say AI is advancing too quickly. 2% say too slowly.
- Society. 40% expect AI to have a negative effect on society over the next 20 years, against 16% who expect a positive one.
- Oversight. 67% have little or no confidence in the US government to regulate AI effectively, and 59% are not confident that US companies will develop and use it responsibly.
The privacy, speed, society and oversight figures are from Pew's report Americans and AI 2026, based on a survey of 5,119 adults in February 2026.
Do people who use AI fear it less?
The surveys do not show that. Between 2024 and 2026 the share of US adults who say they use AI chatbots rose from about a third to 49%, and about a quarter now use one every day. Over the same period, concern did not fall.
Users are not unhappy with the tools themselves. People are far more likely to say chatbots help their productivity (30%) than hurt it (5%). The worry is about something else: what the technology does to work, privacy and society when everyone is using it.
It is possible to find a tool useful and still be afraid of where it leads. A driver can value the car and still want speed limits.
So is the fear rational?
A fear is usually called irrational when it is out of proportion to the actual danger. That is how clinicians separate a phobia from ordinary caution. So the question is not whether people are afraid of AI. It is whether the fear is in proportion to the evidence.
On three counts it holds up reasonably well.
- It is specific. People are not afraid of AI in general. They name jobs, privacy, speed and weak oversight. Specific fears can be checked.
- It is informed. The worry is highest among adults under 30, who use the technology most. Unfamiliarity does not explain it.
- It matches what insiders say. In a World Economic Forum survey of more than 10,000 executives, about 54% expected AI to displace a large number of jobs, and only 12% expected it to raise wages. The public's fear about work is close to the expectation of the people making the decisions. We set out that evidence in what the job forecasts actually say about AI.
On one count it deserves caution. Expecting something is not the same as it happening. That 71% of people expect fewer jobs tells you about the public mood, not about the labour market in 2046. Public expectations about technology have been wrong before, in both directions.
The fairest summary is that fear of AI is a reasonable response to real uncertainty. It becomes unreasonable only when it hardens into certainty, either that catastrophe is coming or that nothing can be done.
When does worry about AI become a problem?
For most people it does not. Worry that leads you to read more carefully, ask questions at work or support better rules is doing its job.
It becomes a problem when it starts to run your life: constant preoccupation, avoiding ordinary technology you need, or physical symptoms of anxiety. Technophobia is not a formal diagnosis in the DSM-5, the standard manual of mental disorders, but clinicians can treat a severe fear of technology in the same way as other specific phobias, often with gradual exposure or cognitive behavioural therapy. If that description fits, a doctor or mental health professional is the right person to talk to. This article is general information, not medical advice.
What can you do with a reasonable fear?
- Name the specific thing. "AI" is too large to think about. "My job", "my data" or "what my children read" are questions you can investigate.
- Check it against evidence. Look for who measured what, and whether the number describes an expectation or an outcome.
- Separate the tool from its deployment. Most harms come from how a system is used and who decides, and those are things people can change.
- Keep using what helps you. Avoiding a technology does not make you safer from its effects on society.
- Say it out loud. Half the country feels the same way. A fear that is shared is the start of a public argument, which is how rules get written.
Fear is information. On the current evidence, the information is that many people have looked closely at this technology and are not yet convinced it is being built for them.