The ‘AI avalanche’ descends on surgery, ChatGPT goes healthcare-specific and more AI news & views not to miss

AI is an icy landslide sweeping through healthcare. Even surgeons must choose between two stark options: Ride the crest of the powder cloud now or dig yourselves out of a deep runout zone later. 

The heads-up comes from a longtime leader in the field who doesn’t speak idly. AI is “forcing healthcare to reimagine the future of surgery,” writes surgery professor and trauma specialist Lenworth Jacobs Jr., MD, MPH, of the University of Connecticut in commentary published Jan. 7. His extracurricular posts over the years have included president of the American Trauma Society, director of the American Board of Surgery and regent of the American College of Surgeons. Other key points straight from Dr. Jacobs:

• For today’s surgeon, AI offers a unique opportunity to transform the way complex care is delivered. “From preoperative diagnostics to intraoperative guidance and postoperative monitoring, AI can integrate multiple clinical domains simultaneously into a coordinated system. Rather than operating in silos, these domains can be woven together to form a seamless, real-time response that can reduce delays and streamline critical decision-making.”
   
• In the OR, robotics paired with AI analytics can allow for more detailed assessments of surgical technique and performance. “Beyond the OR, AI-driven simulation environments enable trainees to practice on lifelike cases, benchmark their performance against national databases and carry those lessons back into clinical care.”
   
• A surgeon’s responsibility is always to the patient—not the algorithm, not the speed, not the financial margins. “Fastest and cheapest are not always best, especially when considering that the priority of patients and their families is the desired outcome. AI is not, and should never be, a replacement for clinical judgment, but if used and developed wisely, it can be used as a tool to enhance surgical skills and aid decision-making.”
   
• Surgeons have an opportunity to be informed stewards of this technology. “[They] must take the time to understand how these systems work, lean into the ways AI can help deliver better care and identify the areas where it is falling short.
   
• Data out are only as good as the data in. “By getting involved now, the surgical community can take the lead in training these tools on everything from broader datasets that incorporate more diverse patient populations to testing limits on diagnostics.”
   
  • The piece, posted by the American College of Surgeons, is applicable to all physicians and, by extension, many if not most healthcare AI stakeholders. Read the rest. 

Remember OpenAI’s humblebrag about ChatGPT answering health-related questions from around the world 40 million times a day? Less than a week later, the company is announcing the addition of a healthcare-specific tab to the LLM’s interface.

Creatively calling the new service ChatGPT Health, OpenAI says users can expect it to deliver a “dedicated healthcare experience that securely brings your health information and ChatGPT’s intelligence together to help you feel more informed, prepared and confident navigating your health.” 

• ChatGPT Health was developed in close collaboration with physicians around the world to provide clear and useful health information, OpenAI says. The company claims it spent two years working with more than 260 physicians from 60 countries. The idea was to get a handle on what makes an answer to a health question helpful, unhelpful or, worst-case scenario, potentially harmful. “This group has now provided feedback on model outputs over 600,000 times across 30 areas of focus,” OpenAI says. “This collaboration has shaped not just what ChatGPT Health can do, but how it responds: how urgently to encourage follow-ups with a clinician, how to communicate clearly without oversimplifying and how to prioritize safety in moments that matter⁠.”
   
  • OpenAI’s online announcement features a slick promotional video with some genuinely inspiring moments. 

While many healthcare AI vendors hype advances in agentic and autonomous AI, their customers and prospects are choosing and using less splashy efficiency boosters.

At the top of the list is ambient listening and transcribing tools. Close to 80% of provider orgs are using the technology. Runners-up, none of which is close, include AI-assisted patient engagement (36%), predictive risk modeling at the population level (35%) and claims adjudication (24%). The figures are from KLAS Research’s latest healthcare AI update, which the firm subtitles “What use cases are adopted the most?” 

• As for industry players, KLAS finds new startups struggling to gain a foothold while incumbents—including giants like Microsoft and Epic—settle in all the deeper. The report is available for downloading here.

Also worthwhile:

• Google, Character.AI to settle suits involving youth suicides and AI chatbots (CNBC) 
   
• Your AI doctor may be working for someone else (Boston Globe) 
   
• AI arms race in utilization review: Promises of efficiency and risks of ‘supercharged’ flaws (Health Affairs)

From HealthExec’s sibling outlets: 

• FDA clears expanded use of AI that helps novice users perform echo exams (Cardiovascular Business)
   
• Current imaging AI systems can be repurposed to address future pandemics, study suggests (Radiology Business)