these authors contributed equally
This is an open-access article distributed under the terms of the Creative Commons Attribution License (
A nationwide survey of youth aged 14 to 24 years on generative artificial intelligence (GAI) found that many youths are wary about the use of GAI in health care, suggesting that health professionals should acknowledge concerns about AI health tools and address them with adolescent patients as they become more pervasive.
Generative artificial intelligence (GAI) has become increasingly prevalent throughout health care, with GAI tools being applied to clinical decision support, medical documentation, patient-provider communication, and more [
This study was approved by the Institutional Review Board of the University of Michigan Medical School. Informed consent was obtained via the internet for all participants. All data were deidentified and protected by a National Institutes of Health Certificate of Confidentiality. Participants received US $1 per week for answering each week’s survey and a US $3 bonus if all questions were answered during the 8- to 12-week phase. US $5 was provided upon enrollment completion, which included a web-based demographic survey.
Study participants were respondents of MyVoice—a nationwide text message survey that collects perspectives from youth aged 14 to 24 years. Five open-ended questions related to general GAI use were texted to participants in March 2024 (
Of 758 eligible youths, 624 (82.3%) responded to at least one question. On average, respondents were aged 20.3 (SD 2.6) years (
Survey response themes are summarized in
Across all survey responses, 10.6% (66/624) of respondents mentioned health applications, with 40.9% (27/66) citing specific health-related applications (eg, creating meal plans, designing exercise regimens, refining medical diagnoses, and improving health care systems) and 63.6% (42/66) expressing concerns. These concerns primarily focused on the need for human input in medical decision-making (12/42, 28.6%) and the importance of avoiding medical errors (8/42, 19%).
Demographic characteristics of 624 youth respondents in the MyVoice survey on generative artificial intelligence.
| Demographic variable | Respondents |
| Age (years), mean (SD) | 20.3 (2.6) |
| Gender identity (respondents: n=624), n (%) | |
| Male | 210 (33.7) |
| Female | 328 (52.6) |
| Transgender | 35 (5.6) |
| Nonbinary/other | 51 (8.2) |
| Race (respondents: n=624), n (%) | |
| White | 361 (57.9) |
| Black | 86 (13.8) |
| Asian | 103 (16.5) |
| Mixed race | 48 (7.7) |
| Other race | 26 (4.2) |
| Hispanic (respondents: n=624), n (%) | |
| Yes | 85 (13.6) |
| No | 539 (86.4) |
| Highest education level (respondents: n=623), n (%) | |
| Less than high school | 123 (19.7) |
| High school graduate | 79 (12.7) |
| Some college or technical school | 236 (37.9) |
| Associate’s or technical degree | 41 (6.6) |
| Bachelor’s degree or higher | 144 (23.1) |
| Region (respondents: n=623), n (%) | |
| Midwest | 203 (32.6) |
| Northeast | 127 (20.4) |
| South | 164 (26.3) |
| West | 129 (20.7) |
| SNAP | |
| Current | 51 (9.0) |
| Previous | 94 (16.7) |
| Never | 419 (74.3) |
aSNAP: Supplemental Nutrition Assistance Program.
Survey themes, responses, and representative quotations from the MyVoice survey on generative artificial intelligence (GAI).
| Theme | Responses | Representative quotes |
| Current patterns of GAI usage | ||
| 159 (27) | “Helping to automate tedious tasks” | |
| 95 (16.1) | “Yes, primarily as a research tool for school” | |
| 27 (4.6) | “I’ve heard of [GAI] being used…to [alert] Healthcare workers to potential outcomes.” | |
| Meal ideas | 9 (33.3) | “Make(s) it easier to find recipes for cooking” |
| Managing disabilities | 4 (12.1) | “I am disabled, it helps me with tasks such as drafting emails (and where)…it matters how you come across.” |
| Exercise | 3 (11.1) | “Basically a free personal trainer” |
| Concerns regarding GAI usage | ||
| 159 (28.8) | “[GAI] can blur the lines between reality and fiction.” | |
| 148 (26.8) | “I am concerned it is going to take jobs.” | |
| 92 (16.6) | “[GAI] can be a crutch, and make people overdependent on technology. It may also be used to stifle creativity…” | |
| 3 (0.5) | “[GAI] can become dangerous when used in medical situations…” | |
| Circumstances in which GAI should not be used | ||
| 180 (31.7) | “Writing academic papers or doing school assignments...” | |
| 103 (18.2) | “It should NEVER be used to create images/videos/…of people without consent” | |
| 115 (20.3) | “AI shouldn’t be used for artistic purposes because art is intrinsically rooted in human emotion and experience” | |
| 40 (7.1) | “The thought of [GAI] in the medical field is unsettling” | |
| Complex scenarios in health care | 5 (0.9) | “Medicine/healthcare - too many components that I wouldn’t trust the machine to handle” |
| Health care because of the need for human input | 12 (2.1) | “Healthcare industry jobs that require human consideration, life or death care I guess” |
| Concern for mistakes or biases in health care settings | 8 (1.4) | “It should not be used to answer serious questions such as legal or medical advice since it can cause fatal errors.” |
| High-stakes situations in health care | 5 (0.9) | “It’s simple for [GAI] to make a simple mistake that can result in someone’s life being on the line.” |
aSurvey responses were coded across all themes. A single response could count toward multiple categories.
Our study found that GAI adoption among American youth (77.2%) is high when compared to the general population (39.5%), which is consistent with previous research [
Interestingly, the use of GAI platforms to address personal health concerns was not mentioned by participants in this study. This finding suggests that self-diagnosis is not currently a primary focus of youth who use GAI tools. Rather, youth more frequently voice concern with safeguarding medical processes against error and ensuring that human input is retained in medical decision-making.
In summary, youth remain wary of GAI applications in health care despite widespread GAI adoption for other use cases. Health care professionals can build trust and rapport with adolescents and young adults by acknowledging and addressing their concerns as GAI use becomes more prevalent in health care. Further research into meaningfully applying GAI to youth health is warranted to guide successful implementation of this novel technology.
This research was funded by the Michigan Institute for Clinical & Health Research, the University of Michigan MCubed program, and the University of Michigan Department of Family Medicine.
The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Data are available from the corresponding author upon reasonable request.
TC and MW have full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. MW and TC were responsible for the concept and design of the study. All authors contributed to the acquisition, analysis, or interpretation of data. CS, MB, MW, TC, and AW drafted the manuscript. MW, TC, and AW critically reviewed the manuscript for important intellectual content. All authors provided administrative, technical, or material support. MW, TC, and AW supervised the study.
None declared.
generative artificial intelligence
Supplemental Nutrition Assistance Program
Probing survey questions with corresponding response rates from the MyVoice survey on generative artificial intelligence.