Industries / Healthcare
AI in Healthcare
Strongest where AI reads images and writes paperwork. Still unproven where it makes the clinical call itself.
- AI-enabled devices on the FDA's list (29 Jun 2026)
- 1,614
- of those devices are radiology tools
- 76%
- clinicians given an AI scribe on one day at Vanderbilt
- 2,400+
Updated daily · newest first
Latest in healthcare AI
Policy
FDA closes comments on how to regulate generative AI medical devices
The FDA’s discussion paper floats a two-axis risk framework, and the American Hospital Association asked it to keep some AI tools outside its oversight.
Policy
Policy
South Korea issues guidelines on using AI chatbots for mental health
The health ministry warns against self-diagnosis and over-reliance, after a survey found most recent users had turned to AI for emotional support.
Policy
Policy
UK accepts all 44 recommendations on regulating AI in healthcare
The UK will move towards monitoring AI medical devices for their whole working life, with draft guidance on device changes due by December 2026.
Policy
Research
IIIT Delhi finds maternal-health models match the average mother, not the individual
An IIIT Delhi study of ten models on maternal-health advice finds they track the average north Indian mother closely but barely shift between different women.
Applications · Research
Technology
Sakana AI’s Japanese-tuned model goes into a clinical evidence search tool
Sakana Namazu is now behind Evidence Finder, a doctor-facing literature search tool, with a 96.4% medical licensing exam score and a very carefully bounded claim attached to it.
Applications · Models
Research
Kyunghyun Cho launches Ortet, a health AI lab with a $500 million commitment
A new New York lab led by Kyunghyun Cho launches with a $500 million commitment from Thoreau and a plan for a patient-centred foundation model, but no product…
Applications · Research
Research
Alibaba’s DAMO Academy publishes two CT-reading cancer models in Nature Medicine and Science
Two models that read routine CT scans for oesophageal cancer and 146 other conditions land in major journals, with strong sensitivity figures that are, for now, Alibaba’s own.
Applications · Research
Adoption
Google’s Pixel Watch can now call an ambulance by itself in Europe
Breathing Emergency Detection escalates on its own, and Google publishes no accuracy figures and names no countries.
Applications
Profile
Reviewed monthly · Sources (9)
Where AI works today
Imaging leads: about three in four AI devices on the FDA’s list are radiology tools. Next comes paperwork: ambient scribes that draft visit notes, and AI for medical coding and insurance denials. Patient intake agents are the newest wave.
Still unproven
Most published results come from single hospitals and small samples, and some are written by the hospitals’ own staff. The FDA says its own device list is not complete. When an AI acts without a clinician, who is liable for a mistake is still unsettled.
Regulation
US: the FDA clears AI devices one by one, and January 2026 guidance eased review for some digital health tools. Texas has required patients to be told when AI is used in their care since 1 Jan 2026. In October 2026 the UK accepted all 44 recommendations of its national commission, including monitoring AI devices for as long as they are in use, and South Korea issued safety guidelines for mental health AI.
Proven deployments
Named organisations, measured results, linked sources only
- 20.1%Vanderbilt University Medical Centerof visit notes used an AI scribe within ten weeks of switching it on for 2,400+ clinicians at once.JAMIA, Feb 2026
- −15%Singapore General Hospitaldocumentation time per visit (5.3 to 4.5 minutes), with 10.6% more eye contact, across 169 visits.JMIR study, Mar 2026
- 48Spanish hospital networkhospitals ran an in-house AI scribe in emergency care for 12 months, then added four more specialties.Int. J. Med. Informatics, 2026
- 77%Mass General Brighamof patients using its AI intake agent reached a physician the same day.OpenLoop roundup, Oct 2026
Who’s putting AI to work
- Mass General BrighamDeployer · intake agent
- Vanderbilt UMCDeployer · system-wide AI scribe
- Houston MethodistDeployer · ambient AI in surgery
- Singapore General HospitalDeployer · in-house AI scribe
- MicrosoftVendor · Dragon Copilot scribe
- GE HealthCareVendor · imaging AI
- AidocVendor · diagnostic imaging AI
- AKASAVendor · autonomous medical coding
What to watch
- Lawsuits in California and Illinois over AI scribes recording visits without clear patient consent.
- How the UK turns lifecycle monitoring of AI devices into actual rules.
- Whether the FDA’s list starts flagging devices built on large language models, as the agency says it plans to.
- Whether autonomous medical coding holds up as insurers question AI-assisted billing.
Sources
- AI-Enabled Medical Device List (FDA)
- FDA has cleared 1,614 AI medical devices (Complizen)
- Enterprise-wide deployment of ambient scribe technology (JAMIA)
- Ambient scribe reduces clinician documentation time (Singapore General Hospital study)
- Ambient AI scribe in emergency care (HealthManagement.org)
- Digital health trends, October 2026 (OpenLoop)
- Healthcare AI News, 7 Oct 2026 (HIStalk)
- AI coverage incl. South Korea mental health AI guidelines (Healthcare IT News)
- FDA landscape and new state AI laws (OTA AI Corner)
Keep exploring
ExplainerWhy AI models hallucinateThe reason every AI-drafted clinical note still needs a human check.
ExplainerWhat is an AI agent?The idea behind patient intake and coding agents.
RoboticsAI Robotics hubSurgical and hospital robots are covered there.
ResearchLandmark papersThe research behind today's medical AI.
EventsAI ConferencesDates for the tech keynotes and research conferences that shape AI.
Healthcare on the timeline
Full timeline →More industries
All industries →
Finance & banking
Manufacturing
Retail Soon
Logistics Soon
Legal Soon
Education Soon
Media & marketing Soon
