Healthcare AI newswatch: Trump-era AI regs, AI sign language, AI dentistry, more

Buzzworthy developments of the past few days.

• The Trump administration will probably play to type when it comes to regulating AI. Which means it’s likely to favor innovation over caution—i.e., “guardrails”—and try to extinguish political fires as they ignite. The forecast, paraphrased here, is from three attorneys with the Cravath law firm in New York City. The three also expect Congress to eschew a broad, EU-style framework and instead target thorny issues and use cases. While the coming shifts take shape, AI sellers and buyers should not “abandon efforts to comply with existing regulations and guidelines adopted under the Biden administration,” the team advises. “And in parallel, businesses should continue work aimed at complying with state laws, which will go into force regardless of the change in Washington.” Bloomberg Law published the opinion piece Feb. 20. Read it here. 
   
• Can you name the third most widespread language in the U.S.? It’s American Sign Language, or ASL. Only English and Spanish are spoken more. Now comes an interactive internet platform to help even more hearing-challenged Americans master the skills. The work is backed by Nvidia in collaboration with the American Society for Deaf Children and the creative shop Hello Monday. The platform is simply called Signs, and it uses an AI tool that lets learners receive real-time feedback on their signing acumen from a 3D avatar. It’s pretty nifty. And here’s something else you may not have known: Most deaf children are born to hearing parents. “Giving family members accessible tools like Signs to start learning ASL early enables them to open an effective communication channel with children as young as 6 months old,” one expert tells Nvidia. “Knowing that professional ASL teachers have validated all the vocabulary on the platform, users can be confident in what they’re learning.” Learn more here. 
   
• One good assistive-technology solution deserves another. At Arizona State University, a graduate student is leveraging AI to help people with visual disabilities. The student, computer science major Kelly Raines, has in mind equipping these individuals with digital tools that let them enjoy the world around them more thoroughly. More safely too. The assistance will come in the form of an AI-powered aide that acts as a visual guide. “As the smart glasses collect images, the wearer can speak questions aloud, asking for more details,” ASU News explains. “A person might ask the AI assistant how many steps of a staircase are ahead, or to identify an object off in the distance or to read the text of a street sign.” To this Raines adds: “We wanted to use the best models to accurately describe the environment, understand spatial reasoning and assist with other important tasks in navigation.” 
   
• Never mind generative AI. Bring on generative health. The physician-scientist-entrepreneur Daniel Kraft, MD, suggests this next step only makes sense given AI’s potential to wring clinical insights from multimodal health data. He’s talking about data gleaned from the usual sources plus genomics, microbiome analysis, metabolomics and social determinants of health. “[O]ur health agents will interact with us based on our age, culture, language, personality type, education level, healthcare goals” and more, Kraft said in a recent talk. Reporting on the session for Forbes, MIT senior fellow and entrepreneur John Werner comments that, with AI making sense of these complex inputs, “we can better understand what is happening inside our bodies daily, enabling more proactive and personalized healthcare approaches. This shift will not only extend our years of optimal health but also minimize the time spent dealing with disease.” Read the rest. 
   
• AI continues making inroads into mental healthcare. It’s now been shown adept at predicting the onset of schizophrenia in behavioral-health patients just from routine clinical data in EMRs. It did less well spotting the approach of bipolar disorder, which can have less definitive symptoms, but showed some promise there too. The research was conducted at Aarhus University Hospital in Denmark. Reporting their findings in JAMA Psychiatry, the team reports that models trained with both structured and text-based predictors performed about the same as models trained with only text-based features. “[T]his underscores the importance of text in clinical prediction modeling within psychiatry,” they write. 
   
• Another researcher keen on AI for mental healthcare comes out and asks if AI will replace human psychiatrists. Henry Miller, MD, a former FDA official, answers: Nah. It won’t do that. AI may, however, “democratize access to high-quality therapy, delivering effective treatment to vast numbers of patients at low cost,” Miller writes in a piece published Feb. 18 by the American Council on Science and Health. “While no AI is yet adequate for independent psychiatric use, it holds the potential to complement and enhance human therapists by providing insights into the nuances of effective therapy and offering detailed analysis of therapy sessions to understand why certain approaches work better than others.” Read the whole thing. 
   
• AI isn’t coming for dentists, either. But it is going to help them. Look for the technology to automate administrative processes in dentistry just as it’s doing across medicine, an industry leader tells Healthcare Innovation. The aim will be to let dentists spend less time fussing with forms and more time focused on teeth. Sound familiar? 
   
• Investors are basking in the hottest period ever for AI. Even so, overall investment in healthcare AI startups per se isn’t topping its peak in 2021. That may change, though. And soon. “[T]here’s no obvious sign that investors are tapping the brakes on investments at the intersection of AI and health,” explains Crunchbase News columnist Joanna Glasner. “We might even see a pickup as more health and biotech startups incorporate AI as a core focus area, given the technology’s rapid advancement and increasing sophistication.” Hear her out.
   
• Recent research in the news: 
   
  • University of Delaware: 3D-printed lung model helps researchers study aerosol deposition in the lungs 
     
  • Oxford: Advances in AI can help prepare the world for the next pandemic, global group of scientists find
     
  • AAAS: AI-assisted diagnosis for immunological disease
     
  • University of Gothenburg (Sweden): Robot faster than physicians at spotting side effects
     
• Funding news of note:
   
  • Six months after raising $29M, Candid Health nabs another $52.5M to ease medical billing
     
  • Data-sharing startup Avandra Imaging raises nearly $18M from top health systems, others
     
  • Vitalchat secures $6M Series A funding to scale virtual nursing and telehealth solutions
     
  • AI healthcare startup OpenEvidence raises funding from Sequoia at $1 billion valuation
     
  • Nexmedis, an AI-powered health information system provider based in Indonesia, secures undisclosed amount of funding
     
• From AIin.Healthcare’s news partners:
   
  • Health Imaging: Risk prediction algorithm slashes number of unnecessary thyroid nodule biopsies
     
  • Cardiovascular Business: Embracing the future: James Min left academia to push for a paradigm shift in preventive cardiology