Whether fearful or excited in nature, discussion around the potential impacts of AI on healthcare remains a very hot topic throughout healthcare facilities across the US. Recently, I was invited by my friends at Technology Crossover Ventures to speak on a panel at the Perelman School of Medicine at the University of Pennsylvania on just this topic – my perspectives of AI in healthcare. I was also joined by industry peers William Boonn of Rad AI and Nick Gossen of Dandelion Health. We debated with a live audience of radiology residents and the takeaways from the conversation left me with both surprising and unsurprising insights.
Nearly seven years ago, computer scientist Geoff Hinton famously quoted that AI is going to replace radiologists within five years. Of course, the concern that high sensitivity and high specificity algorithms would render human radiologists obsolete has not come to fruition. Yet, I figured that the fear likely remained amongst radiology residents, or so I thought. Upon conversing with the residents of UPenn, this was not their first, second, or third fear; in fact, they almost laughed at the idea.
Instead, the radiology residents granted me a different thought-provoking concern to ponder on as we contemplate the future of AI, not only in healthcare, but beyond:
Although we’re constantly surrounded by AI conversations internally at Aidoc and externally, the realist would point out that healthcare AI is still a new technology and has its fair share of skeptics. With that said, there is a mountain of clinical evidence that testifies to the effectiveness of AI in health systems. We believe at Aidoc that its widespread adoption is more a matter of “when” than “if.” The residents at UPenn understood that, and led the conversation in an interesting direction:
Overall, the conversation was more ‘excited’ versus ‘fearful’ in my eyes, and gave me a lot of optimism for the future of AI in radiology as it’s clear the incoming cohort of radiologists are acutely aware of the technological revolution, and in that, are ready to embrace the healthcare AI revolution…with a few caveats.