A new survey reveals that 63% of US adults consider AI-generated health information to be at least somewhat reliable, despite disclaimers accompanying such responses, raising concerns about potential misinformation. Conducted in April 2025 by the Annenberg Public Policy Center with a sample of 1,653 adults, the study found that 75% of those who search online for health information receive answers from AI-generated responses, with 31% reporting these responses always or often provide the information they need. While most respondents (80%) own a smartphone or smartwatch and may share health data with providers, nearly half (49%) express discomfort with healthcare professionals relying on AI tools over their own expertise, highlighting a tension between embracing technological convenience and maintaining trust in human medical judgement.
AI-Generated Health Information Use and Reliability
Approximately two-thirds of United States adults report consulting online resources for health information, with a substantial majority – 65 per cent – having encountered AI-generated responses during these searches. These responses, typically presented as AI Overviews or Copilot Answers, are viewed as providing a needed answer by a considerable proportion of users; 31 per cent report that AI ‘always’ or ‘often’ provides the information they seek, while a further 45 per cent find it helpful ‘sometimes’.
Despite this apparent utility, a majority (63%) express some degree of confidence in the reliability of AI-generated health information, with 55% deeming it ‘somewhat’ reliable and 8% considering it ‘very’ reliable. However, a quarter of online searchers remain unconvinced, classifying the information as generally ‘unreliable’. It is noteworthy that older adults (65+) are less likely to report encountering these AI responses than younger demographics.
While most individuals do not rely solely on AI-generated summaries, with approximately two-thirds continuing their search by following links to other resources, a significant portion of the public remains wary of healthcare professionals adopting AI tools. Almost half of those surveyed (49 per cent) express discomfort with healthcare providers relying on AI rather than their own experience when making decisions about patient care. Nevertheless, a considerable number – 41 per cent – believe their healthcare provider already utilises AI tools to some extent.
The prevalence of smart devices capable of tracking health metrics is high, with 93 per cent of respondents owning a smartphone or smartwatch. A majority of users (52 per cent) are willing to share the summary health data generated by these devices with their healthcare provider, suggesting a degree of openness to integrating personal health tracking into clinical care. Furthermore, a substantial majority of respondents (80 per cent) express confidence in their healthcare providers’ ability to explain health issues and treatments effectively.
Confidence in Health Information Sources
Despite widespread adoption of online health resources, a notable proportion of the public exhibits caution regarding the integration of artificial intelligence into healthcare decision-making. Almost half of those surveyed (49 per cent) express discomfort with healthcare providers relying on AI rather than their own experience when making decisions about patient care. Nevertheless, a considerable number – 41 per cent – believe their healthcare provider already utilises AI tools to some extent.
The prevalence of smart devices capable of tracking health metrics is high, with 93 per cent of respondents owning a smartphone or smartwatch. A majority of users (52 per cent) are willing to share the summary health data generated by these devices with their healthcare provider, suggesting a degree of openness to integrating personal health tracking into clinical care.
Furthermore, a substantial majority of respondents (80 per cent) express confidence in their healthcare providers’ ability to explain health issues and treatments effectively. This contrasts with some reservations regarding AI, indicating continued value placed on the expertise and communication skills of human clinicians.
Survey Methodology
The survey methodology employed a nationally representative panel of 1,653 U.S. adults, recruited and maintained by SSRS, an independent market research firm. The initial panel was established in April 2021, with ongoing replenishment to mitigate attrition and maintain representativeness. A replenishment sample of 360 respondents was added in September 2024, utilising a random probability sampling design. Data collection for this wave of the survey was conducted between April 15th and 28th, 2025.
The margin of error for the survey is ±3.4 percentage points at the 95 per cent confidence level. All figures have been rounded to the nearest whole number, and as a result, totals may not always sum to 100 per cent. It should be noted that combined subcategories may also exhibit minor discrepancies due to rounding practices.
This data represents the 24th wave of a longitudinal study tracking public knowledge, beliefs, and behaviours relating to health issues. The Annenberg Public Policy Center’s team overseeing the survey included research analyst Laura A. Gibson, director of the Annenberg Health Risk Communication Institute, Patrick E. Winpenny, and managing director of survey research, Ken Winneg.
Recent publications stemming from this research include reports on Sudden Infant Death Syndrome (SIDS), mammogram screening guidelines, hypertension awareness, measles and the MMR vaccine, and the risks associated with raw milk and avian influenza. These reports are available through the Annenberg Public Policy Center’s website, along with links to the organisation’s social media channels on Facebook, X, Instagram, and Bluesky.
Online Health Information Seeking Habits
Despite widespread use, a nuanced understanding of how individuals incorporate AI-generated health information into their broader health-seeking behavior is emerging. Approximately (65%) of those who search for health information online continue their search by following links to specific websites or other resources, indicating that AI-generated responses are often used as a starting point, rather than a final answer.
Beyond seeking information, many individuals are actively tracking health data via smart devices. Nearly 6 in 10 (59%) smart device users report utilizing apps on their devices that summarize health information, such as steps per day, sleep time, sleep quality, or heart health. This data is often seen as a potentially valuable resource for healthcare providers, with 52% of smart device users expressing a willingness to share this summary health information with their doctor or other healthcare provider.
However, this willingness to share data does not necessarily translate into complete trust in AI-driven healthcare. While a sizable number of respondents are comfortable with pharmacies sending alerts about their prescriptions, nearly half (49%) remain uncomfortable with healthcare providers relying on AI tools rather than their own experience when making decisions about patient care. This suggests a desire for a balance between leveraging the power of AI and retaining the human element of medical care.
The Annenberg Science and Public Health survey continues to monitor these evolving trends, providing valuable insights into the public’s understanding, beliefs, and behaviors regarding health issues. Recent reports stemming from this research include findings on SIDS, mammogram screenings, hypertension, measles and the MMR vaccine, and the risks associated with raw milk and avian influenza.
AI Response Reliability and User Perception
Beyond seeking information, many individuals are actively tracking health data via smart devices. Nearly 6 in 10 (59%) smart device users report utilizing apps on their devices that summarize health information, such as steps per day, sleep time, sleep quality, or heart health. This data is often seen as a potentially valuable resource for healthcare providers, with 52% of smart device users expressing a willingness to share this summary health information with their doctor or other healthcare provider.
However, this willingness to share data does not necessarily translate into complete trust in AI-driven healthcare. While a sizable number of respondents are comfortable with pharmacies sending alerts about their prescriptions, nearly half (49%) remain uncomfortable with healthcare providers relying on AI tools rather than their own experience when making decisions about patient care. This suggests a desire for a balance between leveraging the power of AI and retaining the human element of medical care.
Healthcare Provider AI Tool Concerns
Despite this apparent utility, a majority – 63 per cent – express some degree of confidence in the reliability of AI-generated health information, with 55 per cent deeming it ‘somewhat’ reliable and 8 per cent considering it ‘very’ reliable. However, a quarter of online searchers remain unconvinced, classifying the information as generally ‘unreliable’. It is noteworthy that older adults (65+) are less likely to report encountering these AI responses than younger demographics.
While most individuals do not rely solely on AI-generated summaries, with approximately two-thirds continuing their search by following links to other resources, a significant portion of the public remains wary of healthcare professionals adopting AI tools. Almost half of those surveyed (49 per cent) express discomfort with healthcare providers relying on AI rather than their own experience when making decisions about patient care. Nevertheless, a considerable number – 41 per cent – believe their healthcare provider already utilises AI tools to some extent.
The prevalence of smart devices capable of tracking health metrics is high, with 93 per cent of respondents owning a smartphone or smartwatch. A majority of users (52 per cent) are willing to share the summary health data generated by these devices with their healthcare provider, suggesting a degree of openness to integrating personal health tracking into clinical care.
Furthermore, a substantial majority of respondents (80 per cent) express confidence in their healthcare providers’ ability to explain health issues and treatments effectively. This contrasts with some reservations regarding AI, indicating continued value placed on the expertise and communication skills of human clinicians.
Despite widespread use, a nuanced understanding of how individuals incorporate AI-generated health information into their broader health-seeking behavior is emerging. Approximately (65%) of those who search for health information online continue their search by following links to specific websites or other resources, indicating that AI-generated responses are often used as a starting point, rather than a final answer.
Beyond seeking information, many individuals are actively tracking health data via smart devices. Nearly 6 in 10 (59%) smart device users report utilizing apps on their devices that summarize health information, such as steps per day, sleep time, sleep quality, or heart health. This data is often seen as a potentially valuable resource for healthcare providers, with 52% of smart device users expressing a willingness to share this summary health information with their doctor or other healthcare provider.
However, this willingness to share data does not necessarily translate into complete trust in AI-driven healthcare. While a sizable number of respondents are comfortable with pharmacies sending alerts about their prescriptions, nearly half (49%) remain uncomfortable with healthcare providers relying on AI tools rather than their own experience when making decisions about patient care. This suggests a desire for a balance between leveraging the power of AI and retaining the human element of medical care.
Smart Device Health Data Tracking
Despite this apparent utility, a majority – 63 per cent – express some degree of confidence in the reliability of AI-generated health information, with 55 per cent deeming it ‘somewhat’ reliable and 8 per cent considering it ‘very’ reliable. However, a quarter of online searchers remain unconvinced, classifying the information as generally ‘unreliable’. It is noteworthy that older adults (65+) are less likely to report encountering these AI responses than younger demographics.
Despite widespread use, a nuanced understanding of how individuals incorporate AI-generated health information into their broader health-seeking behavior is emerging. Approximately (65%) of those who search for health information online continue their search by following links to specific websites or other resources, indicating that AI-generated responses are often used as a starting point, rather than a final answer.
Beyond seeking information, many individuals are actively tracking health data via smart devices. Nearly 6 in 10 (59%) smart device users report utilizing apps on their devices that summarize health information, such as steps per day, sleep time, sleep quality, or heart health. This data is often seen as potentially valuable resource for healthcare providers, with 52% of smart device users expressing a willingness to share this summary health information with their doctor other healthcare provider.
However, this willingness to share does not necessarily translate into complete trust in AI-driven healthcare. While a sizable number respondents are comfortable with pharmacies sending alerts about their prescriptions, nearly half (40% remain uncomfortable with healthcare providers relying on AI tools rather than their own experience when making decisions about patient care. This suggests desire for balance between leveraging the power of AI retaining the human element of medical care.
Public Understanding of Health Risks and Alerts
The prevalence of smart devices capable of tracking health metrics is high, with 93% of respondents owning a smartphone or smartwatch. A majority of users (52%) are willing to share the summary health data generated by these devices with their healthcare provider, suggesting a degree of openness to integrating personal health tracking into clinical care.
Furthermore, a substantial majority of respondents (80%) express confidence in their healthcare providers’ ability to explain health issues and treatments effectively. This contrasts with some reservations regarding AI, indicating continued value placed on the expertise and communication skills of human clinicians.
Despite widespread use, a nuanced understanding of how individuals incorporate AI-generated health information into their broader health-seeking behavior is emerging. Approximately (65%) of those who search for health information online continue their search by following links to specific websites or other resources, indicating that AI-generated responses are often used as a starting point, rather than a final answer.
Beyond seeking information, many individuals are actively tracking health data via smart devices. Nearly 6 in 10 (59%) smart device users report utilizing apps on their devices that summarize health information, such as steps per day, sleep time, sleep quality, or heart health. This data is often seen as a potentially valuable resource for healthcare providers, with 52% of smart device users expressing a willingness to share this summary health information with their doctor or other healthcare provider.
However, this willingness to share data does not necessarily translate into complete trust in AI-driven healthcare. While a sizable number of respondents are comfortable with pharmacies sending alerts about their prescriptions, nearly half (49%) remain uncomfortable with healthcare providers relying on AI tools rather than their own experience when making decisions about patient care. This suggests a desire for a balance between leveraging the power of AI and retaining the human element of medical care.
More information
External Link: Click Here For More
