“There’s an opportunity for a maker movement to help shape AI.” – An interview with Board Member Nathan Wilson
Nathan Wilson, Ph.D., is an expert in artificial intelligence and its deployment for industry and government, and specializes in designing and operating production-grade systems of AI, including within healthcare. As an entrepreneur he co-founded Nara Logics, a Boston-based AI company in 2011 and has since served as its CTO. Dr. Wilson is also a neuroscientist by training whose work has been published in Nature, Nature Neuroscience, Nature Protocols, Proceedings of the National Academy of Sciences, and The Journal of Neuroscience, among others. He earned his Ph.D. from MIT in brain and cognitive sciences, and bachelor of science in neurobiology and cognitive studies and master’s degree in computer science from Cornell University. Deeply interested in the relationship between AI and the mind, he is an experienced contributor to popular articles and interviews about the future of technology, and how we may think.
Lown Institute (LI): Tell us a bit about your background.
Nathan Wilson (NW): I grew up in Ithaca, New York where my mother worked at Cornell University for astronomer and planetary scientist Carl Sagan who was a hometown hero. He was all about the truth, communicating clearly about science, and—perhaps more metaphysically—finding ways to balance the light and darkness, positives and negatives in our world. He was definitely an early inspiration when I got into science.
I studied at Cornell where I received a bachelor’s degree in neurobiology and cognitive studies, and then a master’s degree in computer science. This was in the late 90s when the AI field was really taking off. I went on to get my Ph.D. in brain and cognitive sciences from MIT, where I studied how the cerebral cortex processes, structures, and stores information. I’ve always been very interested in how people arrive at objective truths in complex systems like healthcare, especially when there is so much information—and disinformation—in the public discourse. I went on to co-found Nara Logics for decision-making AI, on the idea that principles from the cerebral cortex—sparse connectivity, associative recall, and inference you can trace—could produce AI that explains itself and learns by connecting across different silos of data.
LI: What led you to make the transition from academia and research in neuroscience to founding an AI startup?
NW: Ithaca is a college town and, for many young people growing up there, the professional goal is to become an academic and that is indeed the route I pursued. I wrote the papers and landed the postdoc, but at the last second I felt an itch that entrepreneurship just might be equivalent or even more freeing than working in a laboratory for the things I wanted to build. In part this was because there just wasn’t the funding structure available in the academy for the direction I was heading in, whereas in the entrepreneurial space there seemed to be an emerging interest in productionalizing AI and how to think about it.
LI: Lown’s board is primarily made up of clinicians and healthcare executives. What is it that attracted you to the board given your different career arc?
NW: I think it was the possibility that we could employ the same type of analysis and approaches used in the AI field to understand the underlying trends in healthcare. This could give people a different way of looking at these systems, rather than simply retreading the same old metrics that may be helpful for the industry but not helpful for the patient or clinician.
I’m also excited by the idea that AI has the potential to help the Institute identify low-value care by providing the analytical data to spot negative trends in much shorter time periods. I was in my patient portal the other day and I was amazed at how it’s getting more and more organized, and now includes data from my older medical history. If we’re able to synthesize data like this at scale, then we’re more likely to both recognize overuse patterns and understand their cause and effect, while at the same time providing a check on our own internalized biases.
LI: We often talk of material divide in this country and, more and more, a digital divide. Do you think AI can lift all boats or will it exacerbate existing inequalities?
NW: No one has a crystal ball, but when I was in college and the internet was first becoming widely available, I wrote an article about the ways the internet was going to lead to just such a digital divide in access. This isn’t what actually came to pass. With the greater advent of AI, I think access to the technologies will be widely distributed to people. However, the bigger questions of systemic inequality will remain, and as with the internet you especially need to watch out for the second-order effects.
LI: How do you envision that we can best employ these AI technologies in clinical settings in a manner that makes things more manageable and transparent?
NW: This question reminds me a bit of the open source movement in technology where there was this idea that if you could give everyone free access to this technology and these software packages, they could shape it into what was actually needed and wanted. I think we’re at a similar point with AI, where there’s an opportunity for a “bottom-up” maker movement to help shape how AI is employed on the front lines. By more intimately knowing patients’ needs and medicine’s current blind spots, I feel that clinicians are positioned to actively guide the development of these tools.
LI: Researchers recently used generative AI to create new synthetic viruses capable of killing bacteria. While engineered to treat antibiotic-resistant infections, the breakthrough sparks massive global biosecurity and regulation concerns. What are your thoughts about this development?
NW: There are definitely some dystopian overtones. These threats are real, in the political sphere and in the technology sphere so we have to be very clear-eyed. That tension is precisely why the application matters. I do think the more we can employ AI to benefit humanity, the more that will reinforce positive uses of the technology. The former CEO of DeepMind has said he wants AI to focus on healthcare because the outcomes are an unambiguous net positive. If AI can help us move the needle on cancer diagnosis and treatment, for example, that’s going to help build the right kind of momentum needed to keep the technology moving in a positive direction.
LI: What do you think AI’s role in healthcare will be in 10 years?
NW: I think AI will be a net elucidator. It’ll provide clarity on what’s happening at the patient level as well as at the aggregate level. There will be misuses that’ll be widely publicized along the way but I think if in 10 years you say, “I’d like to turn off the AI now and go back to the ‘before days’” in healthcare, most people would not agree. It’s going to be so integrated into our day-to-day that we’ll wonder how we made do without it.
LI: As someone who’s seen so much change in their career field in such a short time period, what gives you the confidence that the healthcare system is going to continue to bend towards justice, so to speak?
NW: When I was young, my father would always say you have to give all your energy out and hope that it’ll help. There’s a line from psychologist/philosopher William James, a Harvard professor who is often called the father of American psychology, who said, “Be not afraid of life. Believe that life is worth living, and your belief will help create the fact.” With respect to AI, there’s a kind of paradoxical aspect where if you come to it with a scary sci-fi mindset then we’re actually more likely to fall into a pit of despair. But if you believe that these tools can be used positively, that’s going to subtly, day by day, encourage all of us in a positive direction.
SEE ALSO: