Buzzworthy developments of the past few days.

• Several provider execs testified about healthcare AI on Capitol Hill Nov. 29. Among them was Christopher Longhurst, MD, chief medical and digital officer for UC-San Diego Health. “With thoughtful implementation and careful oversight to ensure equity, transparency and effectiveness,” Longhurst told the House Energy and Commerce Subcommittee, “AI can be transformative for healthcare delivery, improving quality and patient safety, and reducing administrative burdens.” The hearing was recorded and is available for viewing on demand (full runtime: 3 hours, 18 minutes).
   
• AI jobs pay 78% more than other occupations. The pay gap between tech jobs and other occupations has widened by 36% due to AI. And the state with the highest average pay for AI work is Connecticut, at $176,000 per year. The findings are from Bizreport. For additional key findings plus the survey methodology, click here.
   
• EHR giant Epic is training AI to write like the very physicians who use it for drafting answers to patients’ questions. Sumit Rana, the company’s EVP of R&D, explains how it works. “You might have a doctor who says, ‘Dear Bill.’ Another one might say, ‘Hi, Bill. A third one might say just ‘Bill,’ and a fourth one might not say your name at all,” Rana tells HIMSS Media’s Healthcare IT News. “If the writing style does not mirror what the patient is used to from that provider, it can feel very unnatural.” Full interview here.
   
• The UK Biobank is offering whole-genome data drawn from 500,000 volunteers. “This abundance of genomic data is unparalleled,” the organization says, “but what cements it as a defining moment for the future of healthcare is its use in combination with the existing wealth of data UK Biobank has collected over the past 15 years on lifestyle, whole body imaging scans, health information and proteins found in the blood.” The biobank hopes the move will help researchers around the world discover drugs, innovate diagnostic techniques and find cures to any number of diseases. Announcement.
   
• AI programs consume a lot of energy. And 3 of 4 organizations that took the dive into AI admit they were unprepared for the drain. So finds the data storage supplier Pure Storage. Get the rest of the company’s survey findings here.
   
• AI-equipped stethoscopes can help parents monitor kids with asthma right at home. In a proof-of-concept study, the technology as developed by StethoMe (Poznań, Poland) worked well enough that it could help Mom and Dad know when a minor wheeze is stable and fading—or when it’s building to bigger breathing trouble requiring a doctor’s attention. Annals of Family Medicine has published the study in full for free.
   
• Last week a devoted family man had to bolt Thanksgiving dinner to go treat an emergency stroke patient. The incident prompted the dessert deserter—New Jersey neurosurgeon Paul Saphier, MD—to reflect on the strengths and weaknesses of diagnostic AI. “Not only have I been involved in cases where AI has indicated surgery should be performed when objectively there was nothing to operate upon, but the opposite scenario also arises,” he writes. “It is a precarious situation to call a team for emergency surgery when the AI platform instructs no intervention is required.” The author is the husband of breast radiologist and TV personality Nicole Saphier, MD. Read the piece.
   
• From AIin.Healthcare’s news partners:
   
  • Radiology Business: Aidoc invests $30M to create first-of-its-kind artificial intelligence model
  • Cardiovascular Business: Deep learning in cardiovascular imaging: 4 key takeaways for cardiologists
  • Radiology Business: With an uncertain labor market and the rise of AI, RSNA president says radiology faces a ‘10x force’