Artificial Intelligence > Building Ethical AI With Sonya Makhni, M.D., M.S.

Building Ethical AI With Sonya Makhni, M.D., M.S.

By Alison Caldwell, Ph.D. Illustration by Nazario Graziano

Share

Share this:

Why is ethical artificial intelligence (AI) important? Sonya Makhni, M.D., M.S., hospitalist at Mayo Clinic and a medical director of Mayo Clinic Platform, develops systems to mitigate AI bias, safeguard health equity and build novel care delivery models. Mayo Clinic Magazine spoke with Dr. Makhni to learn about her journey and the importance of ethical AI.

You studied neuroscience as an undergraduate and are now a clinical expert in ethical AI. How did your career path unfold?

Both of my parents are internal medicine physicians, so I grew up seeing firsthand how meaningful and fulfilling a career in clinical medicine can be. At the same time, I was always deeply interested in mathematics and technology.

As an undergraduate at Massachusetts Institute of Technology, one formative experience was a course I took in graph theory, focused on its ability to reveal complex patterns in data — an approach that translates to understanding medical data and potentially supporting better clinical decision-making. During medical school, I also completed an M.B.A., which reshaped how I viewed healthcare and the need for systems-level change. After medical school and residency, I completed a master’s degree in bioinformatics to further refine how I might drive some of these systematic changes. In my case, I was interested in using data and technology to improve patient care.

Related Content

A Deeper Dive Into Knowledge Graphs

These digital tools have the potential to transform patient care.

Read More

When I joined Mayo Clinic as a clinician, I also started working with Mayo Clinic Platform. As a Mayo Clinic Platform medical director, I began engaging regularly with startups presenting their AI-driven solutions. Physicians must sort through hundreds of AI tools to determine which one is most suitable to their needs. This led me to ask critical questions: How do we ensure that these tools are responsibly developed and deployed from start to finish? How do we clearly communicate this complex information to healthcare providers so that they can adopt these tools with confidence and trust?

Synapse is your current research and development project. Can you tell us what this tool is and how it will impact care for patients?

Synapse is a clinical decision support tool we are building to help clinicians make sense of complex patient information. From the start, we are designing it with responsible AI principles in mind. That means being clear about how it works and building in monitoring to check for accuracy and bias.

At its core, Synapse uses a knowledge graph, which is a way of mapping relationships between pieces of information. A knowledge graph connects things like symptoms, diagnoses, medications, test results and social factors so we can see how they relate to one another. By organizing information this way, the tool can surface insights that are tailored to the specific patient in front of the clinician, rather than offering a one-size-fits-all answer.

Technology has so much potential to positively impact our patients, providers and workforce if done methodically, ethically and with patient outcomes at the forefront of the conversation.

— SONYA MAKHNI, M.D., M.S.

Instead of just saying “treat atrial fibrillation this way,” Synapse can recognize that this particular person is, say, a 40-year-old pregnant patient with specific comorbidities, and then combine what the guidelines recommend with actual outcome data from similar patients we’ve treated.

Ultimately, the goal is to provide clinicians with tools they can immediately understand and trust while also revealing connections we might have missed, like social determinants or environmental factors that impact health.

What makes Mayo Clinic a unique leader in the conversation around ethical AI?

Physicians and researchers at Mayo Clinic are empowered by the breadth of patient data available to them and the institution’s unparalleled clinical expertise. Through Mayo Clinic Platform, we have access to the largest curated, de-identified clinical dataset in the world, so we have a massive base to pull from when it comes to developing and testing new AI tools and assessing them for bias.

Mayo Clinic clinicians are truly the best of the best, particularly when it comes to treating serious or complex disease. I am grateful that Mayo Clinic encourages thoughtful exploration and innovation in this space. Technology has so much potential to positively impact our patients, providers and workforce if done methodically, ethically and with patient outcomes at the forefront of the conversation. As a global leader in healthcare, it is essential that Mayo Clinic drives these important conversations.