Peer-Reviewed Publication
Nat Med2026August 4, 2026Journal Article
Divergent impacts of explainable AI for dermatological diagnosis on clinicians versus lay people.
Xuhai 'Orson' Xu1, Haoyu Hu2, Haoran Zhang3, Will Ke Wang4, Reina Wang3, Luis R Soenksen3,5, Omar Badri6, Sheharbano Jafry7, Elise Burger8, Lotanna Nwandu9, Apoorva Mehta4, Erik P Duhaime10, Asif Qasim11, Hause Lin3, Janis Karleen Pereira11, Jonathan Hershon12, Paulius Mui13, Alejandro A Gru4, Noémie Elhadad4, Lena Mamykina4, Matthew Groh14, Philipp Tschandl15, Roxana Daneshjou7, Marzyeh Ghassemi16
1Columbia University, New York, NY, USA. xx2489@cumc.columbia.edu.
2Cornell University, Ithaca, NY, USA.
3Massachusetts Institute of Technology, Cambridge, MA, USA.
4Columbia University, New York, NY, USA.
5Johns Hopkins University, Baltimore, MD, USA.
6Northeast Dermatology Associates, Beverly, MA, USA.
7Stanford University, Stanford, CA, USA.
8University of Utah, Salt Lake City, UT, USA.
9OSF HealthCare, Peoria, IL, USA.
10Centaur.AI, Boston, MA, USA.
11MedShr, London, UK.
12Pathway, Montréal, Québec, Canada.
13X=Primary Care, Cambridge, MA, USA.
14Northwestern University, Evanston, IL, USA.
15Medical University of Vienna, Vienna, Austria.
16Massachusetts Institute of Technology, Cambridge, MA, USA. mghassem@mit.edu.
Abstract
Artificial intelligence (AI) is increasingly permeating healthcare, from serving as a physician assistant to powering consumer applications. The opacity of AI algorithms makes the ability of humans to interact with AI algorithms challenging. To overcome this limitation, explainable AI (XAI) provides insight into AI decision-making, but evidence suggests that XAI can paradoxically induce bias in th…
Create a free account to keep reading
Free members get 10 full research views every month across publications, clinical trials, FDA clearances, adverse events, and NIH grants. No credit card required.
Want unlimited research access? See Pro plans