Clinicians are caught in accountability ping-pong: blamed for over-relying on AI, but flagged for variation if they override it. There are hidden value tradeoffs baked into AI. Standardized Model cards are one possible solution.
#MedSky
@tinglongdai.com @proflog.bsky.social @matthewwicker.bsky.social
Check out the latest from the SSRN #blog with includes the top downloaded papers on #AI in Finance for Q4 2025.
Read more: spkl.io/63328AWeMo
#FinanceSky #AcademicChatter #Research
With Dr. Joseph Kvedar and Dr. Daniel Polsky, we published “Policy brief: ambient AI scribes and the coding arms race” in
@natureportfolio.nature.com Digital Medicine on early signals scribes can raise coding intensity:
www.nature.com/articles/s41...
Ambient AI scribes may be the biggest healthcare development of 2025. Even with no mandate and no reimbursement change.
They spread because doctors are drowning in documentation.
But while well-being gains are celebrated, a coding arms race is taking shape:
Special thanks to our speakers:
Natalia Trayanova
@mdredze.bsky.social
@alexisbattle.bsky.social
@tinglongdai.com
Haris Sair
Alex Baras
Rick Carrick
Robert Stevens
Ashley Kieman
Agbessi Amouzou
FDA-cleared #AI medical devices often experience recalls shortly after clearance, particularly those without clinical validation and produced by publicly traded companies. ja.ma/3UFrZqI
Medical AI is as much about forgetting as it is about learning.
Even if data are deleted, models can still remember what they learned. That’s why “machine unlearning” is emerging as a crucial frontier for privacy, regulation, and trust.
New in Health Affairs: www.healthaffairs.org/content/fore...
Figure 2. Recall Counts, Cause, and Affected Units by Commercialization Model
FDA-cleared #AI medical devices often experience recalls shortly after clearance, particularly those without clinical validation and produced by publicly traded companies. ja.ma/4lFOjeA
In our new @jamahealthforum.com paper, we examined timing and drivers of FDA-cleared AI medical device recalls
We show publicly traded companies’ products are far more likely to be recalled. Such recalls often occur within 1st year & with no human testing
Read more: jamanetwork.com/journals/jam...
Thank you for spotlighting our study!
A study finds clinicians rate peers who use generative AI for primary decision-making lower in skill and competence. Framing AI as a verification tool partially mitigates this negative perception but does not eliminate it.
#MedSky #MLSky #MedAI
Grateful to work with a dream team:
Drs. Haiyang Yang, Risa Wolf, Nestoras Mathioudakis, & Amy Knight @jhu.edu @hopkinsmedicine.bsky.social
plus Dr. Yuna Nakayasu, my former MBA/MPH student @johnshopkinssph.bsky.social @jhu.edu, now of McKinsey
Yet, clinicians also saw promise:
Belief that GenAI improves accuracy: 4.30
Institution-customized GenAI viewed even more favorably: 4.96
(7-point scale)
Ratings of the care experience also dropped — from 4.48 ⭐️ to 3.08 ⭐️ (5-star scale).
Framing GenAI as a verification tool helped (clinical skill 4.99, competence 4.94), but the gap remained.
The numbers are striking.
In our randomized experiment of 276 clinicians, a physician who used GenAI as a decision aid tool was rated far lower:
Clinical skill: 3.79 vs 5.93
Overall competence: 3.71 vs 5.99
(7-point scale)
Would you trust a doctor who uses ChatGPT during your visit?
Today, we published a study in @natureportfolio.nature.com Digital Medicine — the first survey of practicing clinicians on how they view peers who use generative AI in medical decision-making.
🔗 nature.com/articles/s41...
With Mariana Socal of @johnshopkinssph.bsky.social & Maqbool Dada of Carey@jhu.edu, in Health Affairs Scholar @healthaffairs.bsky.social: “Prescription for made in America? Tariffs and U.S. drug manufacturing” url: academic.oup.com/healthaffair...
Medical AI is racing ahead; @fda.gov S_FDA oversight must keep pace.
Read the full @ai.nejm.org study here ➡️
bit.ly/fdaai25
Grateful to Branden Lee, Shivam Patel, CrystalFavorito, Sara Sandri, and Maria Rain Jennings, first- and second-year @jhu.edu medical students already shaping medical AI.
Thanks also to Drs Charlotte Haug and Isaac Kohane of @ai.nejm.org for thoughtful guidance.
(6/7)
Safety gap
@fda.gov-cleared AI devices from publicly traded firms are recalled far more often: up to 30 × compared with those from private firms (14.4% vs 1.3% of cleared devices)
Development and commercialization models corrects with patient risk and should guide oversight.
(5/7)
Tech under the hood
Deep learning now powers half of new @fda.gov-cleared devices.
Transparency is improving, yet 62% of all devices still give little or no detail about how their AI works.
(4/7)
AI Clearance curve
Average @fda.gov clearances jumped from 1.4 devices per year in 1995-2014 to 146 per year in 2020-24—a 100-fold surge.
Total count went from 27 in the first 20 years to 729 in the last five.
In-house development drives nearly all growth.
(3/7)
Who is building what?
69% of @fda.gov-cleared AI device manufacturers are private, but public firms make more devices per company.
General Radiology leads with 32%, followed by cardiovascular (18%) and neuropsychiatry (15%).
A booming yet scattered market.
(2/7)
Honored to lead an incredible team of @jhu.edu medical students on the first full look at how @fda.gov-cleared medical AI devices are developed and commercialized, now in @ai.nejm.org.
We tracked 950 AI medical devices. The results may surprise you: bit.ly/fdaai25
(1/7)
"It's not just about higher costs -- it's about readiness." -- @tinglongdai.com, of @jhu.edu, on the threat that tariffs pose to #publichealth preparedness.
www.medpagetoday.com/special-repo...
5/5 As a top exporter leading in diagnostic and lab reagents, a prolonged trade war could render the global scientific supply chain more fragile, costly, and unreliable.
We may be at the onset of a tariff-induced chaos period.
4/5 The U.S. imports billions in lab equipment and reagents annually; many now face 10–54% tariffs.
Even U.S.-built DNA sequencers may rely on German optics or Chinese semiconductors.
High-end tools like precision microscopes aren't produced domestically.
As I told @celestebiever.bsky.social of @nature.com, “These aren’t luxury items. They’re the core infrastructure of modern science.”