Public Trust Remains The Limiting Factor For AI Adoption In U.S. Healthcare
Findings from a national survey of U.S. adults conducted via YouGov (Nov–Dec 2025; n=500).
Artificial intelligence (AI) is rapidly transforming the technical capabilities of medicine, from diagnostics to administrative efficiency. However, widespread adoption relies heavily on public trust and acceptance 1-2. While the technology is well developed, the “social license” for AI in healthcare remains uncertain 3. This blog post summarizes the findings of a national survey designed to gauge the current sentiment of the US general population. The results reveal a public that is cautious, deeply values human interaction over speed, and remains skeptical of AI’s role in critical medical decisions, despite a growing familiarity with AI tools in daily life.
Methods
To understand the prevailing public sentiment regarding AI in healthcare, we conducted a quantitative cross-sectional survey using the YouGov polling platform. 542 adults (age 18+) across the United States completed our survey between November and December 2025 to ultimately generate a representative sample of 500 respondents. Sampling was weighted to be representative of the general US population based on demographics including age, gender, race, education and region.
Results
Public Patterns of AI Usage
Approximately 59% of the respondents reported using an AI tool (ChatGPT, Gemini, etc.) within the past 30 days but there was a notable discrepancy between general AI adoption and healthcare-specific utilization. While 62% believed they could use AI health tools with minimal difficulty, only 36% of the respondents reported actually trying AI tools for healthcare. Common reasons for using AI for healthcare purposes included general curiosity (80%), convenience (62%), and costing less than a clinic, emergency room or urgent care visit (27%).
Concerns About AI Usage in Healthcare
In general, respondents had apprehensions about providers using AI for their care, with 43% of respondents being “uncomfortable” to “very uncomfortable” with their healthcare provider using AI for their medical care. In contrast, only 25% reported feeling “comfortable” or “very comfortable” with the idea. While most respondents shared neutral positions, it is notable that 27% of respondents frankly opposed their doctor using AI for even a second opinion.
Disclosure preferences were also clear with approximately 73% sharing it was “very” or “extremely” important to be informed by their provider when AI would be used for their care. The value of provider oversight was also strong, with 70% stating it was “very” or “extremely” important that the final decision for their health be made by a human rather than an AI tool.
Medical Errors
Respondents were not optimistic that AI would reduce medical errors, with only 30% reporting that AI would make healthcare provider mistakes less frequent. As for who should shoulder the responsibility for such mistakes when AI is implemented, respondents were mixed on where that responsibility should lie, with the majority (56%) stating that the responsibility would fall on a mix of healthcare providers, the company that created the AI program, and the hospital/clinic.
The Primacy of the Human Connection
One of the most profound findings is the public’s prioritization of human interaction over efficiency. Respondents clearly indicated a preference to wait longer to speak with a human healthcare provider rather than engage with a prompt AI-assisted review. Speed was not viewed as a substitute for the perceived safety and understanding provided by a human doctor. This preference is tied to information quality; respondents believe they would feel “more informed” after speaking with a provider than an AI.
Ethical Concerns: Jobs, Privacy, and Triage
Economic anxiety looms large, with 41% of respondents “very to extremely concerned” about AI replacing healthcare jobs. Similar to the prior noted emphasis on disclosure and transparency, privacy also remains an area of concern with 38% of respondents being “very to extremely” worried about it when AI is used.
Additionally, there is strong opposition to AI making emergency triage decisions. The public appears uncomfortable delegating high-stakes prioritization to non-human agents. Opinions on AI’s impact on fairness and bias were mixed, suggesting the public is still forming its consensus on whether AI will function as a great equalizer or an exacerbator of existing disparities 4. The majority of respondents (52%) favored government regulation for the use of AI in healthcare.
The Partisan Divide
Political affiliation emerged as a potential predictor of sentiment, revealing distinct ideological approaches to healthcare AI.
Republicans consistently reported higher comfort and trust levels compared to Democrats. They are more likely to view AI as a tool to enhance speed and are driven significantly by cost factors.
Democrats expressed significantly higher favorability for the regulation of AI in healthcare. They are more focused on accountability, favoring placing liability for medical mistakes on the AI company. Democrats also showed significantly higher concern regarding healthcare job loss.
Conclusion
As AI integration marches towards ubiquity across society, its use in healthcare, an environment that is deeply human, private, at times high risk, requires critical exploration to guide best practices. A key stakeholder that is at risk of being underrepresented in this forward march is the general public, who in their time of need, may become the patient subjected to this technology.
This survey now provides an up to date overview of public perspective on AI implementation and offers specific insights on how they want and expect this technology to impact their healthcare.
The key takeaway remains that the integration of AI into US healthcare faces a significant trust barrier. While the public possesses increasing digital literacy to use these tools, they remain reluctant to substitute human care with algorithmic efficiency. There is a clear desire expressed in the data for the public to be consenting collaborators in this process.
Consistent with prior reports, this data suggests that for AI to be successfully adopted, implementation strategies must prioritize the “human-in-the-loop” model, ensuring that AI supports rather than replaces the provider-patient relationship 5.
Furthermore, differences across political groups suggest that a single policy approach is unlikely to work for everyone. Successful adoption will require addressing the distinct fears of different demographics, assuaging job and liability concerns for some, while demonstrating cost-efficiency and speed for others 6. Further analysis of the survey data is expected to explore specific AI healthcare use cases (chatbots, scheduling agents, screening tools, etc) across these demographics to better inform developers and policy makers.
The current preference for interfacing with human providers over AI supports the continued general trust placed in providers by the public. Consequently, given provider and patient attitudes are often intertwined; gaining public trust with AI in healthcare will likely first require empowering providers to act as trusted intermediaries for this new technology.
References
- Chew HSJ, Achananuparp P. Perceptions and Needs of Artificial Intelligence in Health Care to Increase Adoption: Scoping Review. J Med Internet Res. 2022;24(1):e32939.
- Gundlack J, Thiel C, Negash S, et al. Patients’ Perceptions of Artificial Intelligence Acceptance, Challenges, and Use in Medical Care: Qualitative Study. J Med Internet Res. 2025;27:e70487.
- Busch F, Hoffmann L, Xu L, et al. Multinational Attitudes Toward AI in Health Care and Diagnostics Among Hospital Patients. JAMA Network Open. 2025;8(6):e2514452.
- Osnat B. Patient Perspectives on Artificial Intelligence in Healthcare: A Global Scoping Review of Benefits, Ethical Concerns, and Implementation Strategies. Int J Med Inform. 2025;203:106007.
- Khullar D, Casalino LP, Qian Y, et al. Perspectives of Patients About Artificial Intelligence in Health Care. JAMA Network Open. 2022;5(5):e2210309.
- Heinrichs H, Kies A, Nagel SK, Kiessling F. Physicians’ Attitudes Toward Artificial Intelligence in Medicine: Mixed Methods Survey and Interview Study. J Med Internet Res. 2025;27:e74187.
