A hospital AI chief says medicine has used AI for decades, and that forgetting this is risky. His advice for patients and leaders: judge AI by patient outcomes.
Philip Payne, chief health AI officer at BJC HealthCare and WashU Medicine, says AI is not new to medicine, and acting as if it is could be a "real danger." He made the case at Newsweek's AI Health Summit in New York City. His point: hospitals rushing to adopt the latest tools may ignore lessons that researchers have been building for decades.
"As a PhD-trained informatician, I'm going to say AI is not in its infancy," Payne said. He noted that AI research goes back to the 1950s and 1960s. What changed, in his words, is that the ability to use it at scale has finally caught up with that research.
Payne has long talked about AI in clinics. In 2018, he pointed to computer detection of abnormal heart rhythms as a simple form of AI already in use. Most health systems did not start seriously weighing broad use of more capable AI until late 2022.
Rohit Chandra, chief digital officer at Cleveland Clinic, called today's technology "probably the most immature version that we're going to see." He said organizations need room to try things and accept that some will fail. "We are literally getting on a journey as opposed to buying a finished product," he said.
Bilal Muhsin of medical technology company BD said the biggest problem in medicine is the quality and timeliness of data. Without a steady flow of current information, he said, "we will all fail."
Michelle Stansbury of Houston Methodist Hospital gave an example of what steady data can do. Continuous vital-sign monitoring, such as heart and breathing readings, has helped clinicians spot trouble earlier. In one case, staff in a virtual operations center saw a worrying signal and alerted a bedside nurse. "We've been able to catch things before they happen," she said.
Payne warned against judging AI only by measures that are easy to collect, such as how fast a task gets done. "Outcomes are everything," he said. At BJC and WashU Medicine, he said, no AI project starts without a clear way to measure results for patients. Their joint Center for Health AI aims to move academic research into real hospital settings.
The summit remarks did not include study results or numbers showing how much any of these tools help patients. That gap is worth keeping in mind. If your doctor or hospital uses AI, a few questions are fair:
When a hospital says it uses AI, ask how it proves the tool improves patient outcomes, not just how fast it works.
World models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasksA University of Chicago law professor asked Claude to fix page numbers. It deleted his syllabus and a year of notes instead. Dropbox saved him, not the AI.
Read more: An AI assistant deleted a law professor's year of lecture notes
