Skip to content
DopeSwagYolo

// bio & health

AI in Medicine: Diagnosis, Scribes and Clinical AI News

News and explainers on AI in clinical care: AI diagnosis and medical imaging, AI scribes, health chatbots, FDA-authorized AI devices and hospital adoption.

Latest in AI in Medicine

What we cover in AI in Medicine

This section covers artificial intelligence used in patient care. That means software that reads scans and other medical images, AI scribes that draft clinical notes, and chatbots people consult about symptoms. It also means AI-enabled devices authorized by the US Food and Drug Administration (FDA), and adoption by hospitals and clinics. AI used to design drugs and proteins is covered in the AI in Biotech section.

Reported use has grown quickly. The American Medical Association published a survey of nearly 1,700 physicians in March 2026. In it, 81% said they use AI professionally, more than double the rate in 2023. The survey found they use it most often to summarize research and draft documentation. The FDA says it has authorized more than 1,600 AI-enabled medical devices as of September 2026. Trade publication MedTech Dive reports that the majority are in radiology. Consumer chatbots sit outside that system. The patient safety nonprofit ECRI notes that they are not regulated as medical devices. ECRI ranked chatbot misuse as its top health technology hazard for 2026.

The evidence is uneven. A randomized trial of about 106,000 women in Sweden was published in The Lancet in January 2026. It found that AI-supported mammography screening was no worse than standard reading by two radiologists on interval cancers. Those are cancers diagnosed between screening rounds. It also found that the AI-supported screening had higher sensitivity. A study of 1,298 UK adults in Nature Medicine found that people who used chatbots to assess symptoms in test scenarios did no better than a control group. Articles here report what each study measured and where its limits lie. Everything in this section is general information, not medical advice. Readers should consult a clinician about their own health.

AI in Medicine reference

AI in Medicine: common questions

Can AI diagnose skin cancer?
AI can help detect skin cancer, but it does not replace an examination. A 2024 meta-analysis in npj Digital Medicine found algorithms performed comparably to expert dermatologists on lesion images. Yet only 5.7 percent of the studies it reviewed were prospective. The FDA-authorized DermaSensor device is an aid for physicians. In a study of 1,005 patients its sensitivity was 96 percent and its specificity 21 percent, MedTech Dive reported. Anyone with a suspicious lesion should see a clinician.
Is ChatGPT safe for medical advice?
Evidence so far urges caution. A 2026 Nature Medicine study had 1,298 UK adults work through test scenarios. In it, people using chatbots identified a relevant condition in fewer than 34.5 percent of cases, no better than a control group. A separate study using written cases found ChatGPT Health undertriaged about 52 percent of emergency responses, which covered two conditions. The safety nonprofit ECRI notes chatbots are not regulated as medical devices. For personal health concerns, consult a clinician.
Will AI replace radiologists?
Trial evidence so far shows AI assisting radiologists, not replacing them. In Sweden's MASAI trial of about 106,000 women, at least one radiologist read every exam. The authors said the results do not support replacing health professionals. In September 2026, MedTech Dive reported that Berlin-based Vara received a European CE mark for software that reports mammograms it classifies as normal without a radiologist. Projections published in 2025 found the US radiologist shortage lasting through 2055 unless the workforce grows or imaging use falls.
Will AI replace doctors?
The sources reviewed here do not show that happening as of October 2026. In the American Medical Association's 2026 survey, 81 percent of physicians used AI. Most often they used it to summarize research and draft documentation, while 17 percent used it for assistive diagnosis. MedTech Dive reported in September 2026 that the FDA had not yet authorized a generative AI-enabled device. A Congressional Research Service report had also noted that in June. How clinical roles change in the longer term remains uncertain.

Sources

  1. More than 80% of physicians use AI professionally: AMA survey, American Medical Association
  2. Artificial Intelligence-Enabled Medical Devices, U.S. Food and Drug Administration
  3. AI in medtech is booming. Track new devices here., MedTech Dive
  4. Misuse of AI chatbots tops annual list of health technology hazards, ECRI
  5. Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial, The Lancet
  6. AI support in breast cancer screening: Fewer missed cancer cases, Lund University
  7. Randomized Trial Shows AI-Supported Mammography Improves Sensitivity and Lowers Interval Cancer Rate, The ASCO Post
  8. Reliability of LLMs as medical assistants for the general public: a randomized preregistered study, Nature Medicine (full text at PubMed Central)
  9. ChatGPT Health performance in a structured test of triage recommendations, Nature Medicine
  10. A systematic review and meta-analysis of artificial intelligence versus clinicians for skin cancer diagnosis, npj Digital Medicine (full text at PubMed Central)
  11. Device Classification Under Section 513(f)(2)(De Novo): DermaSensor (DEN230008), U.S. Food and Drug Administration
  12. Dermasensor wins FDA clearance for AI-enabled skin cancer detection device, MedTech Dive
  13. HPI: Radiologist supply, imaging demand will contribute to future workforce shortage, AuntMinnie.com
  14. Vara receives CE mark for first autonomous breast cancer screening, MedTech Dive
  15. 4 questions about the FDA's approach to generative AI, MedTech Dive
  16. FDA Regulation of AI-Enabled Devices (IF13245), Congressional Research Service (archived at EveryCRSReport.com)