Industry Watcher’s Digest

Buzzworthy developments of the past few days.

• Who pays the damages when a physician follows a care recommendation from AI and medical harm results? The question continues to swirl around U.S. healthcare like a dust devil closing in on an inflatable bounce house at a backyard birthday party. Yes, quite a lot like that, if you think about it. In Politico, the nerve-rattling medico-legal scenario gets a fresh thinking-through in an article published March 24. In any number of cases, reporter Daniel Payne points out, doctors may be liable if they use AI and liable if they don’t. And it won’t help that diagnostic AI can be positioned as analogous to GPS: “It’s up to the driver to stay on the road, no matter what instructions are given” by the device, Payne writes. “Making the situation especially thorny for doctors: They also could open themselves up to litigation if they eschew AI.” Several subject matter experts weigh in. Read the whole thing.
   
• Here’s a somewhat related item. This one’s from the ‘You Can’t Keep Everyone Safe From Everything’ file. A growing number of elderly people are using inflatable hip airbags. Many no doubt do so at the suggestion of their doctors. The devices use sensors to trigger inflation upon detecting a fall in progress. They’re pretty good at preventing hip fractures, which is the aim for which they were invented. However, there have also been reports that the bags “have some cultural biases, for example, for people who pray on the ground. The airbags detect that as a fall and keep blowing up.” Reuters includes the example in an article on AI’s potential to “extend healthcare to all” in Europe.
   
• On AI, Amazon is thinking big, bold and personal. As Steve Jobs and friends prophesied a computer in every home back in the ’80s—and then helped make it happen with Apple—so some lead innovators at Amazon are envisioning (and shooting to deliver) a “personal AGI for everyone.” That’s AGI, as in artificial general intelligence. The kind of AI that learns and, in effect, thinks for itself. Picture a bot like that living in your desktop computer and/or smartphone. In healthcare, for example, “Imagine a patient whose physician is automatically consulted by their AGI based on a change in some vital metrics and then care suggestions are brought,” Vishal Sharma, Amazon’s VP for AGI, tells Axios before acknowledging: “There’s more fundamental work that needs to be done.”
   
• Of course, Amazon’s personal AGI planners will have their skeptics. First to snicker will be AI watchers who’ve been airing their doubts over ChatGPT and other iterations of generative AI. The U.K. tabloid Daily Mail rounds up musings from members of this gallery in an article posted March 25. Quoted in the piece are Prof. Gary Marcus of New York University (“We are starting to see signs that generative AI might be a dud”), GenAI vendor exec Dom Couldwell of DataStax (“This area has seen so much hype; it is growing up in public”) and others. More here.
   
• Meanwhile GenAI for healthcare is quietly working some medical wonders. In one, a young adult used the chatbot to correctly put his doctors on the trail of a rare form of diabetes with a genetic origin. The diagnosis “basically took me from like one of these people that was like counting every single thing I eat, to just eating whatever I want,” the patient, Cooper Myers of Texas, tells a TV station in Austin.
   
• This may not portend anything for healthcare AI. Then again, for those thinking about the technology’s future legal wranglings, it may. Tennessee has become the first state to guard musicians against bad actors who would use AI to profit off artists’ faked voices. The statute is called the ELVIS act not only for the Volunteer State’s most famous son but also for Ensuring Likeness, Voice and Image Security. A little creative naming for the old-time rock and roll fans out there. Local coverage here.
   
• In the Caribbean, a tiny island is making a bundle off AI. The island is Anguilla, which registered .ai as its internet domain years back and now finds tech companies clamoring for a piece of that action. As reported in the Spanish newspaper El País, the British Overseas Territory humors some of these suitors and charges them for the privilege. As for the receipts: “Despite the island’s small size and population of around 16,000, domain registration revenue is significant. Estimates indicate registrations could bring in €72 million ($78.3 million) by 2025.” Get the rest.
   
• Research news roundup:
   
  • University of Copenhagen: Physicists develop modeling software to diagnose serious diseases
     
  • Columbia University: GPT-4 for identifying cell types in single cells matches and sometimes outperforms expert methods
     
  • Google Research: Google AI could soon use a person’s cough to diagnose disease
     
  • Penn Medicine: Friend or foe: A closer look at the role of healthcare algorithms in racial and ethnic disparities