For AI to achieve sweeping adoption across U.S. medicine, physicians will need to be assured they won’t be held liable should clinical algorithms make mistakes.

The nation’s doctors also would be more welcoming of the technology if malpractice insurance covered such errors and the privacy of patient data were a lock.

Those are the three top barriers to adoption as identified from a survey recently conducted by the American Medical Association. The group received responses from 1,081 physicians—661 specialists and 420 primary care providers.

The field was about evenly divided between self-described tech adopters (n = 525) and tech avoiders (556). Here are five more highlights from the survey findings. Note: The AMA uses AI for augmented intelligence.

1. Two-thirds of AMA-surveyed physicians perceive an advantage to using tools with AI.

Of these, 22% see a definite advantage and 43% see some advantage. Far fewer see a definite disadvantage (6%) or some disadvantage (5%). Some 25% are neutral, seeing neither advantage nor disadvantage in these tools.

2. Physicians are split about evenly in on the level of excitement and concern over the use of AI.

41% are equally concerned and excited about the technology. 30% are more excited than concerned, and 29% are more concerned than excited.

3. Use cases generating the most positive expectations involve administrative tasks.

Documentation of billing codes, medical charts or visit notes led the chart on excitement, registering with 54% of the field. Not far behind were automation of insurance preauthorization (48%), creation of discharge instructions, care plans and/or progress notes (43%) and generation of chart summaries (42%).

4. Most respondents would prefer to be at least consulted on tech decisions—but one-fifth have no say in these matters.

Half of physicians, 50%, like to be consulted on technology selection and related activities, while 36% wish to be fully responsible. However, 20% currently have no influence on these decisions and 32% merely have their opinions “considered.” On the other hand, 20% have all or most influence and 28% share influence with others.

5. Just over 1 in 5 use ChatGPT in their professional life and a third use it in their personal life.

Some 17% occasionally use ChatGPT in their work and 5% use it often. In their personal lives, 25% use it occasionally and 7% use it often.

Read or download the full survey report here.

Buzzworthy developments of the past few days.

• Healthcare AI is prone to unintentionally discriminate against some patient groups. No surprise there—the challenge has bugged developers for years now. But seldom has the need for multidisciplinary collaboration to mitigate the problem gotten the emphasis it deserves. Now it does, in commentary authored by incoming NIH director Monica Bertagnolli, MD, and published in the December edition of PNAS Nexus. “The ultimate goal of biomedical research is to create a healthier, more resilient society,” she writes. “To achieve this, research must encompass many influences on health. Fundamental science is the bedrock of this research, but discoveries that arise from the laboratory must also be applied in ways that account for a wide variety of factors in the real world.” Read the whole thing.
   
• Claims denials are up. Does AI have something to do with that—something that someone might be sued over at a class-action level? Well, that’s already happening—google Barrows et al. v. Humana Inc., filed this month in Kentucky. Meanwhile CBS News looks at that case as a possible harbinger of suits to come. David Lipschutz, a lawyer who advocates for Medicare patients, tells the network he’s seen “more frequent” and “inappropriate” insurance claim denials this year. He links the rise to the broadening deployment of AI by healthcare payers. “In our experience,” he says, “the use of these algorithmic tools has led to more denials or premature terminations of coverage for things that otherwise would be covered.” Get the rest from CBS.
   
• Kaiser Permanente is ponying up pretty big bucks to fund healthcare AI research. The multistate healthcare powerhouse has selected medical investigators at five organizations to receive up to $750,000 for advancing, specifically, medical decision-making as aided by algorithms. Find out who’s getting the funding and what work they’re getting it for here.
   
• Accelerated cancer research. Improvements in medical productivity. Disease detection through retinal images. These are a few of their favorite things having to do with AI in healthcare. And who are “they?” Harvey Castro, MD, of the University of Texas and colleagues who shared their picks for the top 2023 AI innovations with Fox News this week. Video and article here.
   
• Healthcare AI watchers haven’t heard all that much from nursing executives. So an interview with Kay Burke, RN, MBA, arrives like a cold drink you didn’t know you needed till you sipped it. Burke, CNIO at UCSF Health, took questions from that institution’s magazine. “All the emerging AI trends are in some way alleviating cognitive burden,” she says in the Q&A. “Right now, nurses are typically hunting and pecking to find and input this information. AI’s summary might not be fully accurate, but generative AI isn’t purporting to be.” Read the rest.
   
• The FDA uses AI. So do the NIH, the CDC and the Administration for Strategic Preparedness and Response. Politico gives a quick glimpse into each one’s AI interests. Read it here.
   
• A venture capital firm called Red Cell Partners just raised more than $90 million to help fledgling but promising AI startups. Red Cell’s CEO, 34-year-old Grant Verstandig, tells Forbes the company will concentrate on incubating and nurturing new shops in two realms: defense and healthcare. “Its ambitious goal is to have those companies emerge from stealth in less than two years with contracts already in hand to start generating revenue quickly,” Forbes reports. Read the piece.
   
• AI is getting combined with animatronic devices—specifically robotic assistants and furry “pets”—to offer comfort and companionship to the lonely elderly. If you’re not sure this development isn’t more sad than smile-inducing, you may want to withhold judgment until you’ve seen the cute critters, the healthy encouragements and the results data. Video report from a TV news team in Vermont here.
   
• From AIin.Healthcare’s news partners:
   
  • Health Imaging: Collaboration brings AI-driven imaging standards to healthcare organizations
  • Radiology Business: Use of AI-driven CT workflows in outpatient imaging to better serve patients
  • Health Imaging: If AI is the future, radiology needs to look to the cloud