A 2024 study conducted by the IC² Institute explores what safety-net providers view as the promises and perils of AI in health care.
AI in Health Care: Safety-Net Providers Share Perceived Benefits and Barriers
June 18, 2025
A recent study conducted by the IC² Institute in partnership with the Episcopal Health Foundation (EHF) reveals that safety net providers — practitioners who provide health care to underserved and vulnerable populations — view artificial intelligence as potentially beneficial, but they are deeply concerned about trust issues and are adamant that humans remain at the center of health care delivery.
The IC² study centers on a survey of health care practitioners in rural and urban Texas, plus in-depth interviews with practitioners and thought leaders. The roughly 230 survey respondents included doctors, nurses and other health care professionals. Respondents shared their perspectives on the benefits and dangers of AI, factors that could influence patient and provider trust in AI, and possible barriers to adoption.
Addressing a Research Gap
Though AI holds significant promise for health care, there has been little research into what types of AI-based solutions could be most helpful to safety-net providers, or what unique steps may be required to successfully deploy AI solutions for under-resourced populations. The IC² Institute believes that gap in research is both a problem and a missed opportunity.
“The lack of research perpetuates one of the core concerns about AI — that AI will exclude participation from underserved and rural populations. But it also misses an opportunity to tap into the knowledge of safety-net providers, who grasp the whole range of social, environmental and behavioral factors that impact health incomes — factors which should be considered in building effective health AI systems.” — S. Craig Watkins, IC² Institute Executive Director
A Nuanced View of Health AI
While many providers are optimistic about AI’s potential to improve care, they identify notable barriers to its successful implementation. And they are reluctant to hand over key aspects of care — especially diagnosis, treatment planning and triage — to AI. “AI could be a game changer in health care going forward, but as a resource, not a replacement for humans,” responded one survey participant, a nurse.
Matt Kammer-Kerwick, one of the study’s lead researchers and director of the Bureau of Business Research at the IC² Institute, summed up practitioner perceptions:
“AI is perceived to have significant potential to improve provider workflows and the personalization of care provided to patients. Still, healthcare professionals remain concerned about data integrity, trust, and institutional readiness…Familiarity drives trust, and providers who already trust AI view it as a valuable support system; those with lower levels of trust raised critical questions about governance, privacy, and the risk of exacerbating existing inequities.”
Key Findings
- Potential Benefits: Participants cited the streamlining of administrative tasks, enhanced patient outcomes, and improved diagnostic accuracy as the top benefits of AI. Respondents also highlighted several areas where AI could improve care specifically in under-resourced and/or rural areas — these included remote patient monitoring, diagnostics, and enhanced health literacy.
- Barriers to Adoption: More than half of respondents, 57%, were “neutral” or “not very confident” in their organization’s ability to integrate AI into their workflow. Respondents cited three primary barriers to health AI integration: concerns about data privacy and security; lack of knowledge and insufficient staff training; and lack of funding.
- Hurdles to Patient Acceptance: Though almost half (45%) of respondents believe that their patients would be responsive to AI tools being used in their care; a full 30% believe their patients would not be responsive. Providers noted challenges that may be more acutely felt in underserved areas: disparities in technology access and literacy and cultural differences such as language.
- Varying Levels of Trust in AI Systems: More than half (53%) of respondents report being “neutral” or distrustful of AI. (Those with greater baseline familiarity with AI indicated higher levels of trust in it.) The one condition that practitioners say will most increase trust in AI? That it makes health care better. AI will prove its value when it helps practitioners deliver quicker/better diagnoses, reach more patients, educate patients more effectively.
What’s Next: Defining a Development Approach
Building on the study’s findings, the Institute is recommending several key strategies to ensure that health AI solutions are responsive to the needs of safety-net providers and the patients that they serve. Among the recommendations is a call for statewide educational initiatives to better equip practitioners (and build trust in AI through education); a statewide AI governance conference to create effective guardrails for the ethical use of AI in health care; and the development of robust and locally-responsive evaluation tools and measures. In the coming months, the Institute will continue to engage with stakeholders and explore how AI can be used as a powerful, human-complementary tool to improve health care.
LEARN MORE
View the literature review and full strategy paper (including strategic recommendations) on the IC² Institute website.
MEDIA CONTACT
Emily Spandikow
(512) 475-8900