One Medicare patient waited eighty-three days for their AI reviewer to answer.

Run time: 7:10

In today's episode:

1. Medicare's AI prior-auth pilot denied care, delayed one case 83 days
2. FDA final order: AI radiology software still needs premarket clearance
3. Gemini, ChatGPT, and Claude all fail a heart-attack screening test
4. First AI-discovered drug doses first patient in a Phase 3 trial
5. MIT and Harvard debut real-time AI X-ray-to-CT navigation tool
6. Meta-analysis: AI hits 93% sensitivity on pediatric wrist fractures
7. Medtronic's Hugo surgical robot gets a new FDA-cleared instrument
8. Anthropic: Claude now leads 26% of its own R&D
9. OpenAI discloses six AI models that misbehaved, adds disclosure rules

TL;DR:

• Medicare's AI-driven prior-authorization pilot denied thousands of requests and left one patient waiting 83 days, per new FOIA documents — a real-world case study for outcome-based AI oversight, not just one-time clearance.
• Three general-purpose chatbots (Gemini, ChatGPT, Claude) failed to reliably flag real heart attacks from prehospital ECGs, while the FDA doubled down on requiring full premarket review for AI radiology software updates.
• The first AI-discovered drug (Insilico's rentosertib) entered Phase 3 trials, and Anthropic says Claude now leads 26% of its own R&D — two data points on how fast "AI building AI" is moving, in the clinic and in the lab.

Sources cited:

• STAT News
• PYMNTS
• Heart & Lung
• Insilico Medicine
• Nature
• Emergency Radiology
• MedTech Dive
• Washington Post
• Axios
• FDA's AI-enabled medical device list

Subscribe: YouTube

medAI Times is for educational and informational purposes only. The content does not constitute medical advice, diagnosis, treatment recommendation, or professional clinical guidance. Consult qualified healthcare professionals and refer to official sources before making clinical, research, regulatory, or business decisions.

Google just mapped every possible way a single DNA letter can go wrong.

Run time: 7:23

In today's episode:

1. DeepMind maps all 9 billion possible DNA mutations
2. UK proposes 44 new rules for medical AI
3. Rural AI push meets skepticism from actual providers
4. OpenEvidence restricts its best model over dual-use risk
5. AI spots collapsed lungs in newborns, 87.6% sensitivity
6. AI-measured organ volume linked to kidney cancer survival
7. Claude Fable 5.1 and Mythos 5.1 go live
8. GPT-6 Astra rated "critical" cyber risk by OpenAI

TL;DR:

• DeepMind released the AlphaGenome Atlas, a free public database predicting the biological effect of all 9 billion possible single-letter DNA changes — built with Broad, Harvard, and Boston Children's, with DeepMind's own caveat that it is not clinically validated.
• The UK's National Commission into AI in Healthcare published 44 recommendations after consulting over 12,000 clinicians and patients, proposing a shift from one-time device clearance to lifecycle monitoring plus a patient's right to know when AI is used in their care.
• OpenEvidence launched a four-model family and restricted its top model, Darwin, which the company says scored 100% on the MedQA licensing-exam benchmark, citing dual-use risk.

Sources cited:

• Google DeepMind
• GOV.UK
• OpenEvidence
• STAT News
• Pediatric Radiology
• Annals of Medicine
• Anthropic
• OpenAI

Subscribe: YouTube

medAI Times is for educational and informational purposes only. The content does not constitute medical advice, diagnosis, treatment recommendation, or professional clinical guidance. Consult qualified healthcare professionals and refer to official sources before making clinical, research, regulatory, or business decisions.

Chatbots beat doctors on paper. Then real patients used them and it fell apart.

Run time: 6:50

In today's episode:

1. Chatbot answers rated more empathetic than doctors, preferred 79% of the time
2. GPT-4 alone beat physicians by ~15 points on diagnostic vignettes
3. Google's AMIE matched or beat primary care docs in text consultations
4. Microsoft's system solved 85% of hard NEJM cases vs doctors' 20%
5. Oxford RCT: patients using LLMs did no better than a web search
6. Half of consumer chatbot medical advice carried potential for harm
7. ~1 in 6 US adults now use an AI chatbot for health advice monthly

TL;DR:

• On clean written cases, AI genuinely out-diagnoses doctors — that part is peer-reviewed, not hype.
• On real patients doing their own typing, that edge vanishes: an Oxford RCT (~1,300 people) found LLM users did no better than web search, because patients omit the detail that matters and models mix right advice with wrong.
• What's actually deployed is the opposite of a doctor replacement: clinician tools like OpenEvidence (used by ~65% of US doctors) keep a licensed human between the model and the prescription. No LLM is FDA-cleared to diagnose you solo.

Sources cited:

• JAMA Internal Medicine, Ayers et al.
• JAMA Network Open, Goh et al.
• Nature, Google DeepMind
• Microsoft AI
• Oxford, Nuffield Dept of Primary Care (Payne et al.)
• NBC News on OpenEvidence

Subscribe: YouTube

medAI Times is for educational and informational purposes only. The content does not constitute medical advice, diagnosis, treatment recommendation, or professional clinical guidance. Consult qualified healthcare professionals and refer to official sources before making clinical, research, regulatory, or business decisions.