Projects

Patient Voices on Artificial Intelligence

How can AI tools be integrated into primary care in ways that promote equity, trust, and patient understanding?

Partners & Funders

The Project

AI is rapidly entering health settings, but its integration in primary care raises questions about equity and trust. We’re exploring the perspectives of patients who are historically underserved by healthcare—those most at risk from poorly implemented AI—to drive AI implementation that better meets their needs.

The Outcome

Through qualitative research with patients, we will generate insights and design principles to improve AI tool selection and implementation. Our findings will provide healthcare system leaders with actionable guidance in pursuit of strengthening patient-provider relationships and driving more equitable health outcomes.

Patient Voices on Artificial Intelligence

How can AI tools be integrated into primary care in ways that promote equity, trust, and patient understanding?

Partners & Funders

The Project

AI is rapidly entering health settings, but its integration in primary care raises questions about equity and trust. We’re exploring the perspectives of patients who are historically underserved by healthcare—those most at risk from poorly implemented AI—to drive AI implementation that better meets their needs.

The Outcome

Through qualitative research with patients, we will generate insights and design principles to improve AI tool selection and implementation. Our findings will provide healthcare system leaders with actionable guidance in pursuit of strengthening patient-provider relationships and driving more equitable health outcomes.

Project Background

Artificial Intelligence (AI) technology has been integrated into American health settings in many ways. While AI offers significant potential to improve health outcomes and system efficiencies, its integration into primary care settings also raises questions about equity, trust, consent, and the future of patient/provider relationships.

To ensure that AI tools support equitable and human-centered primary care, it is essential to understand how people experience and perceive these technologies in their healthcare — especially people from communities whose perspectives are not typically represented in tech product development, who have been historically underserved by the medical community, or who are typically less trusting of doctors.

We’re partnering with the Commonwealth Fund to conduct a qualitative design research study around the integration of AI technologies in primary care settings, with a particular focus on patient populations that have been historically underserved by and/or are mistrustful of the US healthcare system. Our research will seek to identify how patients feel about existing integration of AI in primary care, identify language that allows individuals to best understand and consent to such integration, envision ways that AI-supported systems might eliminate gaps in the quality of care patients receive and enable them to be active partners in their care.

Synthesized insights and associated design principles will be delivered to the Commonwealth Fund, where they will be used primarily to inform healthcare system leaders, with secondary audiences including policymakers and AI technologists, to support ethical implementation of AI tools in primary care that serve patients’ needs.

0

Participants engaged

0

Data points collected

What We Found

Inquiry Areas

  • What underlying beliefs, experiences, and values shape historically underserved patients’ preferences around the use of AI in their primary care?
  • How do historically underserved patients perceive and weigh the potential benefits and harms of the use of AI in their primary care?
  • What communication strategies can help make the use of AI in primary care legible and subject to patient agency?

To explore these questions, we spoke with historically underserved communities in New York City and Southwest Virginia, prioritizing low-income individuals from populations historically underserved by or mistrustful of healthcare systems. In New York, participants were predominantly people of color; in Virginia, they were predominantly white.

We recruited individuals who had attended at least one primary care visit in the past year so they could reflect on recent experiences with AI in healthcare. We also sought diversity in insurance coverage and familiarity with AI tools. 

Using semi-structured interviews, we explored our inquiry areas and captured reactions to specific AI applications and clinical contexts. We also used design stimuli, a set of scenario cards representing potential uses of AI, to tease out participants’ perceptions and tradeoffs. 

To complement our primary research, we compiled academic literature, professional journals, and reputable innovation news sources to fill gaps and provide additional context.

What We Heard
The videos below provide a preliminary overview of key moments from the field before synthesis or meaning-making.

Get in Touch

Interested in partnering with us?

Join Us

Follow Along

Enter your email below to subscribe to our occasional newsletter.

Wondering what you’ve missed?


Check out our

Social Media

Donate

The Public Policy Lab is a tax-exempt

501(c)(3) nonprofit organization.