Americans are increasingly concerned about AI exacerbating mental health problems

Two new surveys from YouGov about artificial intelligence (AI) and its usage explore AI's intersection with mental health. Americans are increasingly concerned about AI exacerbating mental health problems, and many wouldn’t be comfortable using it in place of a professional therapist. However, younger Americans are more likely to say they would be comfortable working with an AI therapist and more likely to say they could form a deep emotional bond with an AI chatbot.

Americans are more likely now than they were in June 2025 to say they are very concerned about the possibility of AI resulting in the exacerbation of mental health problems (43% vs. 35%).

Among adults under 30, 45% say they are very concerned about this, up from 29% in June. There has also been an increase among 45- to 64-year-olds, to 50% from 37%.

Women are more likely than men to say they are very concerned about the possibility of AI exacerbating mental health problems (48% vs. 38%).

Two-thirds (66%) of Americans are uncomfortable with the idea of working with an AI therapist if they were seeking therapy rather than having an in-person visit, including 49% who say they would be very uncomfortable. 23% say they would be very (8%) or somewhat (16%) comfortable working with an AI therapist.

Adults under 30 are about twice as likely as older Americans to say they would be comfortable working with an AI therapist (37% vs. 20%).

In August 2025, fewer adults under 30 (23%) and 65 and older (13%) said they would be comfortable working with an AI therapist.

Republicans are slightly more likely than Democrats to say they would be comfortable working with an AI therapist (30% vs. 23%).

The majority (55%) of Americans say AI is worse at conducting therapy than are people who do it professionally, including 32% who say it is much worse than people. 14% say AI is about equal to people when it comes to conducting therapy, and 10% say it is better than people.

A decade from now, will AI be better than human therapists? 50% think it will still be worse than people at conducting therapy, 14% think it will be about equal, and 15% think it will be much better than people.

Women and men are about equally likely to say AI is worse at conducting therapy than people (57% vs. 53%), but women are more likely than men to think AI will be worse than people at conducting therapy in 10 years (57% vs. 43%).

Only 6% of Americans say they have used AI for therapy; 17% say they haven’t done so but would consider it. 7% say they know someone else who has used AI for therapy.

Far more Americans say they’ve used AI to get advice on a personal matter (25%) or get health advice (30%). 14% say they have engaged it in conversation for companionship.

One in 10 Americans (10%) say they definitely (2%) or probably (8%) could form a deep emotional bond with an AI chatbot companion. Far more say they probably could not (18%) or definitely could not (59%) form a deep emotional bond with an AI chatbot. 1% of Americans say they already have formed a deep emotional bond with an AI chatbot companion.

16% of adults under 45 say they could definitely or probably form a deep emotional bond with an AI chatbot companion; 5% of older Americans say the same.

Republicans are slightly more likely than Democrats to say they definitely or probably could form a deep emotional bond with an AI chatbot companion (15% vs. 9%).

6% of Americans think they could definitely (2%) or probably (3%) have a romantic relationship with an AI chatbot companion. 8% say they probably could not do this, and 78% say they definitely could not.

10% of adults under 30 think they could definitely or probably have a romantic relationship with an AI chatbot companion.

Related:

• Americans are increasingly likely to say AI will negatively affect society
• Many Americans think AI companies should be able to limit how the U.S. military uses their tools
• Most Americans say AI will reduce the number of jobs in the U.S.

See the results for the March 31 - April 2, 2026 YouGov survey and the April 1 - 5, 2026 YouGov survey.

— Carl Bialik contributed to this article

Methodology: This article includes results from two 2026 online surveys conducted among 1,099 U.S. adult citizens on March 31 - April 2 and among 1,082 U.S. adult citizens on April 1 - 5. The following methodological description is true for each survey. Respondents were selected from YouGov’s opt-in panel to be representative of adult U.S. citizens. The sample was weighted according to gender, age, race, education, 2024 presidential vote, 2020 election turnout and presidential vote, baseline party identification, and current voter registration status. 2024 presidential vote, at time of weighting, was estimated to be 48% Harris and 50% Trump. Demographic weighting targets come from the 2019 American Community Survey. Baseline party identification is the respondent’s most recent answer given around November 8, 2024, and is weighted to the estimated distribution at that time (31% Democratic, 33% Republican). The margin of error for the overall sample is approximately 4 percentage points.

Image: Getty (Moor Studio)

What do you really think about President Trump, American politics in general, and everything else? Share your reality, join the YouGov panel, and get paid to share your thoughts. Sign up here.