What if Leonardo da Vinci had AI?
Dispatches from the Visualizing Science Symposium: The Impact of Art on Scientific Discovery
For nearly two years now, I’ve been writing these posts that could be summarized simply by one name: Leonardo da Vinci. He perfectly embodies everything I’ve been trying to say. His art inspired his science as much as his science inspired his art and most importantly, he made no distinction between them. I could save us all a lot of time by just pointing to him and saying, “There. Do that.” But I keep writing to explore how to recapture this notion in today’s increasingly specialized world. So last month I drove from Boston to New York City for a one-day symposium on Science, Art, and AI, and found that I’m not the only one trying to do this.
The symposium was held at the Italian Academy for Advanced Studies at Columbia University in a stone-walled room known as the Teatro, with floor-to-ceiling glass doors flanked by red velvet curtains and opening onto a garden courtyard. It felt like stepping back into the Renaissance.
What better place to try to recapture the spirit of Leonardo da Vinci as we envision where AI will fit in? Which got me thinking.
When our Dad was still around, he would gather the family on the porch of the rural Indiana home he and our Mom had built as their empty nest. Under a blanket of stars, we would sit and talk—not about our jobs, or politics, but about Big Ideas. He called them our What-if-Sparticus-had-a-Piper-Cub talks, after an old Saturday Night Live skit. And so as I sat under chandeliers and listened to talks ranging from the history of medical illustration to what AI holds for its future, I couldn’t help but wonder: what if Leonardo DaVinci had AI?
Widely considered the first to create precise and detailed anatomical drawings, AI likely wouldn’t have helped him in this regard. He eventually could have trained a model on his own drawings, but as former Association of Medical Illustrators president Jill Gregory demonstrated in her talk, AI still gets anatomy wrong, and not just its disturbingly botched hands. As for his paintings, da Vinci again invented brand new techniques based on his own intense observations of Nature and optics.1
What about his inventions then? DaVinci didn’t formally study math or engineering until much later in life, so his inventions were based solely on his obsessive observation, insatiable curiosity, and his superior ability to draw what was in his mind. The gentle vegetarian invented many weapons of war but only one was made in his time—the wheel lock that creates the spark that ignites a firearm. Most of his inventions were hundreds of years before his time. The technology needed to bring his designs to fruition hadn’t been invented yet. If he’d had AI, might helicopters have been flying around in the 15th century?
I decided to ask AI, which found it to be a “fascinating thought”. It pointed out that if da Vinci had AI, he could have performed complex calculations quickly and accurately, analyzed existing knowledge and patterns to make connections between different fields of study, organized and categorized his vast notebooks of observations and ideas, iterated and improved on his designs, and created predictive models, helping him to anticipate the behavior of his inventions under various conditions.
But back in the Teatro, the afternoon panel of mostly computer science professors from Columbia stressed that because it tends to be inaccurate, off-the-shelf AI shouldn’t be used to get factual information.2 Panelist Gaël McGill, on the faculty at Harvard Medical School and founder of the revered scientific visualization company Digizyme, further challenged us to consider the ethics of AI, its environmental impact, and what it takes literally and figuratively from artists. All agreed that, for now, AI is best used as a creative partner, to get us out of ruts. Da Vinci didn’t need a creative partner.
Another speaker, Harvard professor of bioengineering and applied physics Kit Parker, created a lab that might be the closest thing I’ve seen to da Vinci-level creativity coupled with the technical know-how to bring ideas to fruition. When Parker became increasingly disillusioned by what he calls “the maladaptive way we build scientists,”—a reductionist approach he sees as beating all the creativity out of them—he ripped out part of his lab and built studio space to get artists and scientists to work together. He hosts chefs, fashion designers, rappers, and all manner of interdisciplinary artists. Parker only cares that he is importing the most creative people in the world.
It’s not that the artists dream up ideas and the scientists execute on them. Rather, the proximity of artists and scientists creates an environment in which the scientists themselves become more creative. During his talk, Parker pointed out that da Vinci could never have gotten tenure at Harvard because they wouldn’t have been able to figure out which department he belonged in, nor would he have been hired to begin with because they couldn’t have put a search committee together.
It has been argued that da Vinci’s genius relied on his being unbiased by knowledge. Likewise, Parker’s artists are unburdened by the specter of everything in bioengineering that hasn’t worked in the past. Perhaps this is why, on the use of AI, Parker says it just wouldn’t work. He believes that innovation requires the juxtaposition of the “young developing scientist and the creative cowboy artist”. Both seem to reside within Parker. With his southern accent, he recounted his realization that BBQ brisket is the opposite of tissue engineering. As he explored creating new materials for wound healing, he talked to pit masters and ended up inventing a new smoker. As a lieutenant colonel in the US Army Reserve who completed two tours of duty in Afghanistan, he made the creative leap to realize that wound dressing and sustainable food packaging are both fundamentally the same problem: “Moisture in, funk out.”
But Parker’s big dream is to create an artificial heart. One could easily imagine the da Vinci sketches of such a device, given his obsessive interest in fluid dynamics in general and blood flow in particular. His prediction on the mechanics of heart valves opposed the prevailing wisdom of his time. Just over a decade ago, he was shown to be right.3
Not long before that, Parker took the first of two fateful trips to the New England Aquarium with his daughter and was struck by how the motion of jellyfish reminded him of beating hearts. He and his team built tiny jellyfish with rat heart cells and silicone rubber.4 Inspiration struck at the stingray touch tank five years later as he observed the way the stingray moved in response to his daughter’s touch, so they built a light-guided miniature stingray out of similar materials.5 Like da Vinci’s, these inventions came from keen observation and intense curiosity. Thankfully for Parker, he has the requisite training and lives in a time when the technological leaps are within reach. An artificial heart might still seem to exist in the realm of da Vinci’s 15th-century design of a flying machine, but each innovation brings it closer to bioengineers’ reach. Let’s hope Parker keeps his aquarium membership active.
Most of what I know about da Vinci, and what appears here, came from Walter Isaacson’s excellent biography of him entitled “Leonardo da Vinci”.
It does hold much promise for science however when models are trained on accurate and useful information and not just… all of the internet.
From “The Heart of Leonardo”, written by practicing cardiac clinician and anatomist Dr. Francis C. Wells. Fun fact: the foreword was written by His Royal Highness Prince (now King) Charles.
J.C. Nawroth, et al. A tissue-engineered jellyfish with biomimetic propulsion. Nature Biotechnology, 2012, 30, 792–797.
S.-J. Park, et al. Phototactic guidance of a tissue-engineered soft-robotic ray. Science, 2016, 353 (6295), 158-162.