Public health and AI wearables | Partner news | Whom to sue, invite into AI governance and altogether avoid

News and views you ought to know about: 

• The question of legal responsibility for AI gone wrong in healthcare remains unsettled. The only sure thing is a lot of finger pointing. A decisive precedent may not come until regulation takes shape, stakeholders adapt and the technology matures. That’s according to an overview of AI in healthcare with particular interest in the legal and ethical angles. The piece is on offer from two subject-matter experts at the global law firm A&O Shearman. “Accountability will be a particular challenge given the complexity of the value chain and the interplay of different regulatory regimes,” writes attorney Jieni Ji of the firm’s Hong Kong office with David Egan, assistant general counsel for global digital and privacy issues at the pharma giant GSK. “It will be important for all stakeholders to engage in continued dialogue to ensure that legal frameworks keep pace with technological developments and that patient safety remains a central focus.” Other passages worth a neon streak from a digital highlighter: 
   
  • “From the perspective of legislation by sovereign states, the legal landscape for AI in healthcare is still in its infancy. Many countries are currently relying on existing technology-neutral laws, such as data protection and equality laws, as well as industry standards, to address AI-related matters. Additionally, some nations are taking proactive steps to develop approaches to address issues arising from AI technologies.”
     
  • “Leading AI developers are also setting up in-house AI ethics policies and processes, including independent ethics board and review committee, to ensure safe and ethical in AI research. These frameworks are crucial while the international landscape of legally binding regulations continues to mature.”
     
  • “[T]eams reviewing AI systems should consist of stakeholders representing a broad range of expertise and disciplines to ensure comprehensive oversight. For example, this may include professionals with backgrounds in healthcare, medical technology, legal and compliance, cybersecurity, ethics and other relevant fields as well as patient interest groups. By bringing together diverse perspectives, the complexities and ethical considerations of AI in healthcare can be better addressed, fostering trust and accountability.”
     
    • Read the whole thing. 
       
• The American Medical Association seconds that last notion. For healthcare organizations diving in with AI, “it is indeed crucial to form robust working groups that include people who represent a broad group of departments,” the AMA states in an article posted July 31. “Physicians, nurses and other clinical staff—as well as representatives from the operational, financial, legal, compliance, technology, data science, pharmacy and patient experience branches of the organization—need to be part of the decisions being made when implementing AI.” 
   
  • The item quotes AMA chief medical information officer Margaret Lozovatsky, MD. “[E]very clinician and every administrative user of these tools has different needs, so it’s important to involve those stakeholders in the conversation to understand what the needs are and to ensure that technology is actually solving problems rather than creating them,” she says. “I’m a pediatrician. My needs are different than a cardiologist or a dermatologist. That’s why it’s important to have the right voices in the conversation.”
     
    • Read the rest. 
       
• Google’s healthcare AI suite recently made a goof during a test that could have been dangerous in the real world. Med-Gemini verbally blended two parts of the central nervous system, commenting on a section of the brain that doesn’t exist. Google people shrugged off the blunder as a mere misspelling. Medical AI skeptics used the episode as reason to distrust AI in clinical settings. “The problem with these typos or other hallucinations is I don’t trust humans to review them,” Maulin Shah, chief medical information officer at 51-hospital Providence health system in Washington State tells The Verge. “These things propagate.”
   
  • The outlet also spoke with radiologist Judy Gichoya, MD, head of Emory University’s Healthcare Al Innovation and Translational Informatics lab. It’s in the very nature of large-scale AI models, she says, to make things up rather than admitting it doesn’t know something. This tendency, Gichoya adds, is “a big, big problem for high-stakes domains like medicine.”
     
    • Read the piece.
       
• In the eyes of investors, the healthcare innovation economy is complicated. Fundraising dollars are down—yet, overall, private markets are looking better than they have in years. That’s the word from Silicon Valley Bank, which just released a midyear report. “AI may not change the face of clinical care, but it’s coming into its own behind the scenes as a valuable tool for tackling long-standing industry pain points,” the authors write. The report shows U.S. healthcare venture-capital fundraising totaled $3B in the first half of 2025. The figure represents a plummet from 2024 and may be healthcare’s worst year in more than a decade, SVB notes. And yet. “Trailing 12 month healthtech AI deal activity has grown ~2x since 2022 and accounted for nearly a third of all healthcare investment in the first half of 2025,” the report states. “Healthcare-specific AI models, pure software drug-discovery platforms and myriad new administrative tools have furnished healthtech with attractive investment opportunities.” Download an expanded report preview here. 
   
• ‘Religion is text and story and ritual. All of that applies here.’ Applies where? To the Rationalists, a growing Berkeley, Calif., community whose members seem obsessed with AI. The quote is from a Catholic nun and theology professor with the Franciscan order, which has been observing (and praying for) our changing world since the 1200s. Sister Ilia Delio and others, including insiders, were asked for their take on these capital-R Rationalists by The New York Times. 
   
  • A Harvard chaplain goes further than the nun in raising a flag over the group. Noting that some of the most powerful minds in Silicon Valley align their thinking on AI with the Rationalists, Greg Epstein—who wrote a book discussing technology as a new religion—asks: “What do cultish and fundamentalist religions often do? They get people to ignore their common sense about problems in the here and now in order to focus their attention on some fantastical future.”
     
  • Rationalist Alex K. Chen says it’s not like that at all. “It’s a place where serendipity can happen,” he tells the Times. “Some people liken it to a college campus or the MIT Media Lab.” 
     
    • Article here.
       
• From AIin.Healthcare’s sibling outlets:
   
  • Cardiovascular Business: AI company raises $55M with help from leading health systems 

The ambient AI playbook: Lessons from two leading health systems

At the recent CompassionIT Summit, leaders from Akron Children’s Hospital and Denver Health shared powerful lessons from rolling out ambient documentation to over 1,500 clinicians. Their biggest takeaway? Stories, not stats, drive adoption. Whether it was a heartfelt testimonial that swayed an entire department or a 60-second Nabla demo that eliminated training anxiety, the common thread was simplicity, authenticity, and clinician-centered design. Read more about the way these health systems are navigating ambient AI implementation: https://dhinsights.org/news/the-ambient-ai-playbook-lessons-from-two-leading-health-systems

Wearable health devices outfitted with AI seem poised to transition from consumer novelties to an integral layer of U.S. public-health intelligence. If certain conditions are met—standards consolidation, equitable subsidization, algorithm auditability, workforce training—the result could be earlier outbreak detection, finer chronic-disease surveillance and more precise resource allocation.

That’s the verbatim take of researchers in the U.S., U.K. and Nigeria who reviewed the relevant scientific literature, concentrating on the U.S. 

Lead author Musa Olayinka Hanafi of the University of Houston and colleagues had their work published July 23 in the Journal of Medical Science, Biology and Chemistry. Here are some key excerpts from their paper. 

GENERAL OBSERVATIONS

1. Wearable devices began as pedometers for weekend joggers.

Since then, they have “evolved into networked biosensors capable of revealing how whole communities breathe, sleep, move and falter in real time,” Hanafi and co-authors write. More: 

‘When AI transforms those raw pulses into patterns, public health surveillance gains precious lead time and health system planners acquire a dynamic map of looming demand.’ 

2. Promise is inseparable from peril. 

“Sensor physics can magnify bias, adoption skews toward the affluent, plumbing is still fragile, and clinicians will disengage if data reach them unfiltered,” the authors point out. “These are not technical footnotes but existential constraints; fail to manage them and the enterprise collapses into noise or mistrust.” 

‘The road forward is therefore both prosaic and profound. It runs through standards bodies and reimbursement schedules, through equity grants and algorithm audits, through nurse training sessions and privacy charters.’ 

3. If these seemingly insignificant components align, the result will be a remarkable health infrastructure.

Such a system could “detect the initial signs of an outbreak, predict ICU bed requirements and direct preventive resources toward sedentary areas,” the authors note. “And it could conduct all these activities without compromising individual autonomy.”

‘In that future, wearable data are neither a gimmick nor an afterthought. They are a shared public utility, as fundamental to collective well-being as clean water or reliable weather forecasts.’

FUTURE DIRECTIONS

1. Next-generation wearables are edging closer to full vital-sign parity with intensive-care telemetry.

Continuous, cuff-free blood-pressure monitors for adults have secured preliminary 510(k) clearances [from the FDA], “hinting that ubiquitous, non-invasive hemodynamics will soon inform both bedside care and population dashboards.”

‘At the silicon layer, bespoke edge-AI chips promise milliwatt inference, bringing anomaly detection entirely on-device and shrinking privacy risk.’

2. Data from those chips will not live in isolation.

Smart-city pilots are already wiring wearable feeds into environmental and mobility grids, Hanafi and co-researchers write, “letting planners overlay particulate surges or heat domes on live physiology.”

‘Digital-twin initiatives take a step further by connecting each citizen’s sensor trace to a virtual avatar, which enables the testing of policy scenarios before investing money or lives.’ 

3. Generative-AI health coaches will interpret those streams in plain language.

These coaches—now in beta at tech giants and startups—“nudge users toward sleep or diet changes and feeding aggregated adherence metrics back to public health.”  

‘If standards, subsidies and audit frameworks mature apace, the coming decade could see public health officials consult a [virtual] physiologic twin of the nation as routinely as meteorologists check Doppler radar—anticipating, not merely recording, the next wave of need.’

The paper is available in full for free. (Click “Download PDF.”)