AC OBSERVATION

Dr. Anthony Chang

 

“The real voyage of discovery consists not in seeking new landscapes, but in having new eyes.”
— Marcel Proust

As I prepared to deliver a keynote at Spark25 in Istanbul, focusing on healthcare innovation and entrepreneurship in the age of artificial intelligence, I found myself reflecting on my own AI journey. And I’ve come to this conclusion: we need thousands of clinicians who are "trilingual" in clinical medicine, AI, and entrepreneurship. Not someday. Now.

Liftoff: leaving the comfort zone

Now is the best time to leave the gravity of your comfort zone and explore distant knowledge planets. With generative and agentic AI readily available, access to knowledge once reserved for domain experts is within reach for every clinician.

AI is our spaceship, freeing us from the familiar and launching us into the vast unknown. But it’s not just about jumping into tech. The farther you travel from your core discipline, the bigger the potential dividend, especially when creativity and human wisdom remain central.

Even though I had a strong foundation in math and statistics, my distant planet was AI - first explored decades ago through chaos theory research with Dr. Bill Norwood. For others, it may be health economics, business strategy, or digital design. The important thing is to start navigating.

The journey: transdisciplinary and exponential

Innovation flourishes at intersections. Like the mathematical operation of convolution, where two functions combine to yield a third, new and transformative. The same applies when medicine meets data science, or when clinicians collaborate with engineers and entrepreneurs.

AI makes it easier to acquire new domain knowledge. Begin with a short online course (like the ABAIM program), lead or join a clinical AI project, or co-write a paper. You’ll multiply your knowledge when you surround yourself with collaborators from diverse backgrounds.

One of my proudest transdisciplinary moments was working with the renowned adult heart surgeon, Dr. Michael DeBakey to design an implantable pediatric ventricular assist device (VAD) with his team of biomedical engineers and device experts as well as healthcare entrepreneurs.

Splashdown: coming back changed

Like astronauts who return to Earth forever changed by the “overview effect,” clinicians immersed in AI often experience a shift in perspective. But beware of tech-centrism - many startups fail because they chase innovation instead of solving a real healthcare problem.

Startups should consider building digital twins of their business strategies, inspired by NASA's Apollo 13 mission, to simulate outcomes and stress-test decisions. I've had my own ups and downs in this space: one exit, two failures, and four ongoing projects. All humbling. All educational.

The future belongs to convolutional clinicians - healthcare professionals who combine diagnostic empathy with data science and entrepreneurial vision. These clinicians won't just adapt, they'll lead healthcare into a more resilient, intelligent, and human future.

Convolutional intelligence: a call to action

AI won't replace clinicians, but clinicians who use AI will replace those who don't. To remain relevant and effective, clinicians must embrace this new “language” alongside medicine and entrepreneurship. Not as an optional skill, but as a professional imperative.

Let’s make “failure is not an option” our guiding motto, not just for space missions, but for the mission of healthcare transformation.

Join the conversation at AIMed25

This call to action will be a central theme at AIMed25, the world’s leading clinician-led conference on AI in healthcare. With more than 1,000 participants, three specialty tracks (pediatrics, education, mental health), and unique offerings like attendee-curated sessions, ABAIM certification, and an abstract competition, AIMed25 is more than a conference. It's a movement.

Whether you’re a student, data scientist, clinician, or CEO, this is the place to get inspired, get educated, and get connected.

See you in San Diego.

Note. These were initially all my original thoughts but I prompted GPT 4o with these commentaries to yield (convolve) a final version that now includes a few of the GPT4o suggestions. I believe that it is still best to have your human thoughts first, then partner with your favorite large language model to yield the final “convoluted” version to take advantage of the best of both human and artificial intelligence.