As a pre-med and computer science major who worked at AT&T and Lucent Technologies during my summers, I’ve always been interested in how technology can help to uproot health inequities. During my career, I’ve provided input into developing health information systems in health care and public health and have shepherded data equity and data governance initiatives in the HIV arena. While AI as a technology has been with us for several decades now (I remember taking my first AI course in the late 90s), it’s been amazing to see the rapid growth of generative AI recently.
A recent blog series by Stephen Norris highlights the increasingly prominent role of AI in health care including embedded biases, an example of how to subvert these biases, and AI’s potential for addressing health equity.
Recently, I learned of Zainab Garba-Sani’s A.C.C.E.S.S. AI model for community-centered AI implementation in health care. The model helps to ensure the voices, priorities, and needs of marginalized communities are centered in the design, development, implementation, and use of AI tools in health care and guide the process which is also informed by multiple stakeholders, or careholders, a term I prefer to say. This is an exciting model that will be a North Star for community-engaged AI research as we work to develop AI-enabled tools that help to eliminate health inequities.
The potential for GenAI to advance health equity is enormous. However, we need funders to invest in these efforts urgently given how fast AI is evolving and the potential harm that could be incurred if we don’t get ahead of embedded biases in the technology.
The image for this blogpost is what Canva’s image AI generator created when prompted with “health equity” and “artificial intelligence”. It created a picture of Black people in white lab coats with tables walking outside in different directions. GenAI still has a way to go and the people who could be potentially most harmed by it need to be at the table to ensure its equitable and inclusive development.