Americans aren’t OK with generative healthcare AI that was developed without physicians and is sold by vendors without track records. Oh, and their own doctors had better be upfront about using the technology—even if it’s tried and true.  

So reports Wolters Kluwer Health, which bases the conclusions on responses from a representative sampling of 1,000 U.S. adults who were surveyed about six weeks ago.

Here are five other key findings from the executive summary of the survey report.

1. Feelings about GenAI for healthcare are all over the emotional map.

• 45% of Americans say their initial feeling toward GenAI is curiosity.
• 44% report concern as their first reaction toward the technology.
• 17% say they feel scared about the use of GenAI in healthcare.

2. Among those who initially feel concerned, the hesitancy tends to come from uncertainty over where GenAI may be getting its information.

• 86% of Americans see a problem in not knowing the source of GenAI training data and/or its methods of testing and validation.
• 82% are uneasy over the possibility that GenAI training and testing data came from raw internet searches.

3. Confidence levels are split down the middle.

• Just over half the U.S. population (52%) would feel secure in the results if they knew their own provider was using GenAI to assist in their care. A little under half (48%), not so much.

4. Want a hand in changing perceptions? Prioritize transparency and education.

• 80% of Americans want to know whether or not their personal provider will rely on GenAI for an upcoming exam or diagnosis.
• If so, 63% will be rattled even if the GenAI product came from an established source, was developed by medical professionals and is constantly updated.

5. Doctor-patient relationships will continue to matter, regardless of GenAI’s adoption in healthcare.

• 63% of Americans would change providers upon learning their current provider was using GenAI.
• Gen X and Baby Boomers would be less likely than Gen Z and Millennials to move to another provider not using GenAI (58% and 55% vs. 71% and 69%).
• The generational gap is likely due to stronger doctor-patient relationships among older Americans, as nearly half of adults under 30 don’t have a primary care provider.

In a news release, Wolters Kluwer Health executive Greg Samios comments that it will take “responsible testing as well as understanding the importance of using the most current, highly vetted content developed by real medical experts to build acceptance of this new technology in clinical settings.”

Buzzworthy developments of the past few days.

• Pope Francis is pleading with the nations of the world to come together around a binding treaty regulating the development and use of AI “in its many forms.” The Vatican released the written message Dec. 14, ahead of Francis’s presenting it live during the World Day of Peace on Jan. 1. Reuters reports the Pope will urge AI developers and users to prevent algorithms from “replacing human values”—and to beware of a “technological dictatorship” that might threaten human life. More here.
   
• Allina Health, Boston Children’s Hospital, CVS Health and 25 other big names in healthcare have signed onto the Biden Administration’s pledge to advance healthcare AI only responsibly. The latest 28 follow the lead of Amazon, Google, Meta, Microsoft, OpenAI, Anthropic and Inflection from last summer. Read the White House’s Dec. 14 blog post here.
   
• Meanwhile Accenture, Augmedix, Deloitte, HCA Healthcare and BenchSci are among the players working with Google Cloud to bring the latter’s healthcare-specific foundation model to life in the U.S. Calling the model MedLM, Google says it’s “the future of our medical generative AI.” Details here.
   
• HHS has finalized its rule for advancing health IT interoperability and AI transparency. The move follows the release of the rule’s proposal last April. Called HTI-1, the rule addresses patient access, interoperability and standards, including enhanced information-blocking requirements, interoperability-focused reporting metrics and USCDI Version 3. Details here.
   
• Only a quarter of healthcare executives interviewed by KLAS Research have implemented generative AI solutions to date. However, well more than half—58%—say their organization is likely to implement or purchase a solution over the next year. More findings plus a link to order the full report here.
   
• The Defense Advanced Research Project Agency, aka DARPA, has awarded the University of Maryland School of Medicine $7.3 million to innovate technology for triaging trauma patients. The work will include tapping machine learning to quickly predict levels of need for lifesaving interventions, especially in mass casualty events and other emergencies in which medical teams are overwhelmed by sheer patient counts. Details.
   
• Need help rekindling a romantic relationship? There’s an app for that. It proved effective in a peer-reviewed scientific study, with 8 of 10 participants reporting “improved and healthier” couplehoods in just a month of usage. More here.
   
• From AIinHealthcare’s news partners:
   
  • Health Imaging: FDA has now cleared 700 AI healthcare algorithms, more than 76% in radiology
  • Radiology Business: South Korean imaging AI vendor Lunit to acquire rival for $193M
  • Cardiovascular Business: Cardiology now has more than 100 FDA cleared AI algorithms; experts say that is just the beginning
  • Radiology Business: What’s the missing link to get radiologists using AI?
  • Cardiovascular Business: 5 key criteria for AI prediction models in cardiology
  • Health Imaging: Virtual reality ‘gaming’ used to plan delicate cancer surgeries