Prefer to listen? Here's the narrated briefing on this area.
Listen to Healthcare and medicine
Audio streaming, downloadable PDFs and certificates come with Plus and Pro.
Evidence & Reach
DeepMind's AlphaFold predicted the structure of every known protein and earned its creators the 2024 Nobel Prize in Chemistry — research that previously took years per protein now takes minutes. The FDA has cleared 700+ AI-enabled medical devices, primarily in radiology and cardiology imaging where AI consistently matches or exceeds specialist accuracy for early-stage cancer detection. AI-powered clinical scribes (Abridge, Nuance DAX, Suki) are returning roughly two hours per day to primary-care physicians by handling documentation in real time. In 2026 the systemic layer caught up: the US Medicare ACCESS program opened the first federal payment lane for AI care agents — a reimbursement milestone that moves clinical AI from pilot toward standard-of-care economics — and a peer-reviewed Science study found OpenAI's o1 model out-diagnosed two attending physicians on emergency-room triage cases. In October 2026 the evidence moved from case files to real patients: a study in The Lancet, led by Google and Beth Israel Deaconess Medical Center, had 98 people consult Google's AMIE chatbot from home before an urgent primary-care visit, with no conversation stopped on safety grounds, the doctors' final diagnosis on AMIE's list in 90 percent of cases, and clinicians saying its summaries helped them prepare for three in four visits. AMIE made no treatment decisions, and the authors call for larger trials.
Sources:AlphaFold (DeepMind / Nobel Prize 2024)FDA AI/ML-enabled devices listNEJM AI Clinical StudiesMedicare's AI-built payment model (TechCrunch, 2026)First AI-designed vaccine enters trials (ITV, 2026)AMIE clinical study in The Lancet (Google, 2026)
A Specific Story
A primary-care physician at Kaiser Permanente, using Abridge during patient visits, finishes notes by 5:30pm instead of taking them home at 9pm. Multiply that by 70,000 US primary-care physicians and you have a real answer to the physician-burnout crisis — and patients getting more eye contact during their appointments, less typing.
What's Next
The next wave: AI-designed antibiotics (the first novel structural class in 60 years, identified by MIT researchers using machine learning, is heading toward clinical trials), AI-augmented surgical robotics with sub-millimeter precision, and predictive risk models that flag deterioration in ICU patients hours before traditional vital-signs alarms. AI has also begun designing therapeutics outright rather than only analyzing them — the University of Cambridge started human trials of the first vaccine whose core component was designed entirely by AI, and in one striking case GPT-5 cracked a three-year immunology puzzle from an unpublished lab dataset.
Recent Top AI Stories showing progress in this area.
Google's AMIE chatbot interviewed real urgent-care patients, and doctors found it useful
A study in The Lancet, led by Google and Beth Israel Deaconess Medical Center in Boston, tested Google's AMIE diagnostic chatbot with 98 real patients, who chatted with it from home before an urgent primary care visit while physicians watched in real time. No conversation had to be stopped under the study's safety rules. AMIE's list of possible diagnoses included the doctors' final diagnosis in 90 percent of cases, and clinicians said its summaries helped them prepare for three in four visits. AMIE made no treatment decisions, and the authors say larger trials are needed before patient-facing AI is used at scale.
Alibaba open-sources a CT model that beat 23 of 26 radiologists
Alibaba's DAMO Academy published RADAR in the journal Science and released it: a vision-language model trained on more than 400,000 contrast-enhanced abdominal CT exams paired with their clinical reports, with no manual annotation. It flags 146 findings across 18 organs in a single pass, including pancreatic and liver cancers, and averaged an area under the curve of 0.913 across nearly 40,000 real-world exams. Against 26 radiologists it outperformed 23, and radiologists working with it raised detection sensitivity by 10 percent while taking 30 percent less time. The code is Apache 2.0, but the model weights on Hugging Face carry a non-commercial Creative Commons license.
An AI tool helped doctors predict which lung cancer patients respond to immunotherapy
The I3LUNG project, a study across six centers in Italy, Germany, Greece, Israel, Spain and the United States, reported in Nature Medicine that AI models built from 2,396 patients with advanced non-small cell lung cancer predicted immunotherapy outcomes better than every standard clinical biomarker, including the PD-L1 test doctors rely on. When 20 physicians reviewed 100 real cases with the tool, their accuracy in spotting responders rose substantially, with the largest gains among doctors who were not lung cancer specialists. This phase looked back at past patients; a prospective trial of more than 2,000 is enrolling.
DeepMind maps the predicted effect of every single-letter DNA change
Google DeepMind released AlphaGenome Atlas, running its AlphaGenome model across all nine billion possible single-letter changes in the human genome and publishing the predictions as a one-petabyte dataset, more than 30 times the size of the AlphaFold database. It is free for non-commercial use, with commercial access on Google Cloud to follow. Early collaborators, including the Broad Institute, Boston Children's Hospital and Memorial Sloan Kettering, report finding 22 percent more non-coding genetic associations and pinpointing a variant in the DNM1 gene tied to a severe childhood epilepsy.
Google researchers estimate body fat and metabolic risk from an ordinary phone photo
Google's PhotoScan model estimates body composition — including the visceral fat ratios that a scale and a tape measure cannot see — from standard two-dimensional smartphone pictures. It was pre-trained on more than 35,000 UK Biobank participants with clinical scan ground truth, fine-tuned on 677 adults photographed with real phones, and validated against a 132-person longitudinal trial. It beat smartwatch bioimpedance on body fat percentage and improved insulin-resistance classification well past demographics alone. Google calls it a research prototype, not a product.