Industry Watcher’s Digest

Buzzworthy developments of the past few days.

• A month after naming its first chief AI officer, Children’s National Hospital in D.C. is building out a dedicated AI operation. According to Washington Business Journal, the institution is adding “dozens” of people who will report up to the new CAIO, Alda Mizaku. The newspaper reports that Children’s National is creating roles for data managers, engineers, product managers and others. The outlet notes the recruitment drive comes on the heels of a $1 million donation from Amazon Web Services.  
   
• The Sam Altman-Arianna Huffington ‘AI health coach’ has attracted detractors. One is Jathan Sadowski, a senior research fellow at Monash University in Australia. Even if the coaching platform, Thrive AI Health, somehow manages to avoid the usual pitfalls—bias, hallucinations and other fumbles—it will “still miss the mark because the idea of hyper-personalization is based on a flawed theory of how change happens,” Sadowski writes. He makes his case over at The Conversation.
   
• He’s got to admit it’s getting better. Then again, it couldn’t get much worse. The antecedent of the impersonal pronoun is healthcare technology. The holder of the opinion is Nirav Shah, MD, MPH, a senior scholar at Stanford. “Healthcare is not productive. The more technology we get, the less productive we become,” Shah said at a VentureBeat event last week. “It used to take me 45 minutes to admit a patient in the paper-based world, and now, thanks to electronic health records, it takes me an hour and 45 minutes. I’m a glorified data entry clerk.” While his present is tinged with disappointment, he has high hopes for healthcare technology’s AI-aided future. Read VentureBeat’s own coverage.
   
• Seconding that present-tense emotion are two legal pros. “While the potential of AI technology is exciting in its transformative potential, we are well served to remember that not all innovations make life simpler,” write Harry Nelson and Yehuda Hausman of the healthcare-specialized firm Nelson Hardiman. “While AI-empowered ‘conveyor-belt’ healthcare brings the promise of new levels of efficiency, it also brings risks. Without adequate human supervision and oversight, minor issues and errors within the new frameworks can easily escalate by several orders of magnitude.” Read the rest.
   
• Armed with certain AI models, amateur inventors could engineer serious biological threats. With this frightening scenario in mind, researchers at Los Alamos National Laboratory are putting heads together with peers at OpenAI. Among other things, they’re interested in learning how bad actors could use multimodal frontier models for nefarious purposes involving biological threats. The two started the project earlier this year. The next phase involves testing experts’ handiwork with ChatGPT-4o for completing real-world tasks—like introducing foreign genetic material into host organisms—and assessing unspecified “emerging biological risks.” Learn more from Los Alamos Lab here and OpenAI here.
   
• Meanwhile OpenAI is telling the world how it could get from AI to AGI in 5 simple steps. Saying its eyes remain fixed on attaining artificial general intelligence, OpenAI has hinted to Bloomberg that AI models capable of performing “a range of tasks across different domains without human input” are still a ways off but already on the whiteboard. Tom’s Guide breaks it down.
   
• A large financial software company is laying off 1,800 workers but planning to hire around the same number. Driving the doubletake-inducing switcheroo is a strategy by the company, Intuit, to infuse its products as well as it processes with AI. “For example, [artificial intelligence] is helping experts with AI-supported answers and explanations, and matching and routing customers to the right expert at the right time, tailoring customer-specific needs to expert profiles and availability,” a company spokesperson tells the San Diego Union-Tribune.
   
• ChatGPT is the most commonly used AI work tool in the world. Hot on its heels are Canva AI Suite and Google Gemini. Quartz has posted a slide show of the top 10 in this category as of May.
   
• Recent research roundup:
   
  • University of Cambridge: Artificial intelligence outperforms clinical tests at predicting progress of Alzheimer’s disease
     
  • Hebrew University of Jerusalem: Beyond algorithms: The role of human empathy in AI-enhanced therapy
     
  • MIT: When to trust an AI model
     
• Funding news of note:
   
  • Regard Raises $61M to close the clinical insights gap in healthcare with leading AI solution
     
  • Wall Street analysts bullish on genetic testing firm Tempus AI
     
• From AIin.Healthcare’s news partners:
   
  • Health Imaging: AI tool accurately completes leg length measurements up to 87% faster than human readers
     
  • Cardiovascular Business: AI creates accurate 4D heart scans in seconds