Industry Watcher’s Digest

Buzzworthy developments of the past few days.

• AI has tremendous potential in healthcare—potential to get things wrong. The warning comes from the immediate past president of the American Medical Association, Jesse Ehrenfeld, MD, MPH. Of course, the technology is also stocked with expedients to help clinicians get things right, and fast, allowing time to check its guidance before acting on it. The trick is to treat algorithms more like aides than experts—and to make each one show its credentials. “We must demand transparency,” says Ehrenfeld, who’s a practicing anesthesiologist as well as a medical informaticist. If he finds himself in a surgery suite in which AI is controlling a patient’s ventilator, he needs to know: “How do I hit the ‘off’ switch?”
   
• AI can’t save healthcare from its own systemic shortcomings. Avoiding the cliché about putting a Band-Aid on a severed artery, tech ethicist Alex John London, PhD, suggests AI won’t improve healthcare delivery until U.S. healthcare undergoes broad, structural change. Deficits that foundational “are not going to be changed by doing fancy work on your dataset,” he told attendees at a recent healthcare AI conference. “To really make use of AI and get all the value out of AI in healthcare, we have to change health systems, the data that we generate, our ability to learn, the way we deliver healthcare and who’s included in our systems.” London is director of Carnegie Mellon’s Center for Ethics and Policy and chief ethicist at the university’s Block Center for Technology and Society. He’s also co-editor of a textbook titled Ethical Issues in Modern Medicine. Get the rest from GeekWire.
   
• Say generative AI is overhyped without coming right out and saying generative AI is overhyped. The MIT robotics pioneer and serial tech entrepreneur Rodney Brooks does pretty much that in an interview with TechCrunch. “When a human sees an AI system perform a task, they immediately generalize it to things that are similar and make an estimate of the competence of the AI system,” Brooks says. “And they’re usually very over-optimistic, and that’s because they use a model of a person’s performance on a task.” What many miss, he politely points out, is that task performance is no measure of general competence.
   
• OK, now come right out and say it. “AI is the hot buzzword, and there is a lot of hype.” That’s from Felice Verduyn-van Weegen of the life sciences division at EQT, a global investment organization. Her interest is in AI startups worth investing in. She tells Private Equity International her firm would “tread carefully when looking at an investment opportunity where AI is the only key differentiator.”
   
• On the bright side, AI might save many people from having a stroke. The hope is expressed by an individual with a vested interest in the proposition. That doesn’t mean the scenario is not worth watching for. “I think in the very near future we will be able to look at a person’s electrocardiogram (ECG) results and, even if they’re not symptomatic, we can [have AI] look at their records and assign a risk ratio” predicting likelihood of stroke, the optimist, InfoBionic.AI CEO Stuart Long, tells Medical Device Network. “We will be able to say with a high degree of confidence that someone is going to develop atrial fibrillation and we can start treatments today that would help offset that. That is somewhere [healthcare] AI is going to help the most.”
   
• Healthcare AI is even of interest to the Armed Forces Communications & Electronics Association International. A July 1 article in the AFCEA outlet Signal looks at AI’s potential for promoting preventive care and improving care access. Belinda Seto, deputy director at the NIH’s Office of Data Science Strategy, says data sharing has never been about technology. “It’s about culture,” Seto explains. “The idea of generating data spanning many years of research, and only sharing it much later, is no longer acceptable. It’s about common good; it’s about community good.”
   
• The American College of Radiology is certifying rad practices that prove their AI prowess. It makes sense that medical imaging pros would be at the forefront of this kind of thing, as radiology was one of the first medical specialties to adopt clinical AI (not least for assistance with image interpretation). Plus it’s still further ahead with the technology than most others. AIin.Healthcare’s sister news site Radiology Business had the story last week. This week it’s interesting to see Politico, of all press outfits, is on it too.
   
• Vicariously experiencing a physician’s tête-à-têtes with ChatGPT is worth the price of the book that chronicles the interactions. That would be ChatGPT, MD by Robert Pearl, MD, with “co-author” ChatGPT. The book recommendation comes from Robert C. Smith, MD, distinguished professor of medicine and psychiatry at Michigan State University. Reviewing the book for Psychology Today, Smith seems not to disagree with a key Pearl takeaway. “[T]he medical revolution needed to overturn the medical-industrial complex,” reviewer Smith writes, “will be mediated by AI in conjunction with two precipitants: societal displeasure with healthcare and the economic catastrophe current practices promise.”
   
• AI can bring out the dark side in people who are willing to pay to engage with it. An attractive, 20-something influencer found this out the hard way. Her mistake: tapping the technology to create her interactive digital double so she could maximize her billable time online as a (well-meaning) loneliness reliever. Her unwanted result: An awful lot of fans, mostly men, got the wrong idea about what the unreal twin was open to discussing. Some wanted to delve into “disturbing fantasies.” At least one customer made like a stalker, setting up a “shrine-like photo wall” of the woman—with the bot’s encouragement. The Los Angeles Times gives the woman’s story without going into lurid details.
   
• Funding news of note:
   
  • Zoi Capital founders and industry veterans announce $200M fund targeting AI technologies with practical applications for healthcare
     
  • SoftBank Group establishes $93M joint venture with Tempus AI to advance Japan’s healthcare through the utilization of medical data and AI
     
  • AI primary care platform startup K Health raises $50M in equity funding round
     
  • $20M gift from Dwight and Dian Diercks to accelerate Mayo Clinic’s strategy for transformational healthcare AI and innovation
     
• From AIin.Healthcare’s news partners:
   
  • Cardiovascular Business: FDA clears new AI algorithm for 1-year AFib risk
     
  • Health Imaging: Bayer and Rad AI expand access to AI and radiology workflow solutions with new collaboration
     
  • Cardiovascular Business: Dual approvals: AliveCor gains FDA clearance for advanced AI model, handheld ECG system