PERSPECTIVE article
Front. Digit. Health
Sec. Health Technology Implementation
Volume 7 - 2025 | doi: 10.3389/fdgth.2025.1665961
This article is part of the Research TopicAdvances in Artificial Intelligence Transforming the Medical and Healthcare SectorsView all 14 articles
'Do we need explainable AI in ophthalmology, or just accurate AI?'
Provisionally accepted
- Humanitas University, Rozzano, Italy
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Artificial intelligence (AI) is reshaping ophthalmology by paving the way for disease diagnosis, prognostication, and personalized treatment planning to occur automatically. Currently, clinical AI models prioritize high accuracy, bringing about noteworthy performance in detecting conditions such as diabetic retinopathy and age-related macular degeneration. However, the issue of explainability-interpreting AI decisions into understandable and transparent language-is controversial. This perspective outlines the balance between accuracy and explainability in ophthalmic AI, arguing that while accuracy is an absolute necessity, explainability is needed to enable clinician trust, bias detection, and permit ethical deployment. We review current explainability techniques employed for retinal imaging AI, their limitations, and the future in integrating explainable AI into routine ophthalmic practice. Finally, we urge a balanced approach where both explainability and accuracy drive the future of ophthalmological AI tools.
Keywords: Artifici al Intelligence, Explainability, accuracy, Ophthalmology (MeSH), deep learning
Received: 14 Jul 2025; Accepted: 25 Jul 2025.
Copyright: © 2025 Buonamassa and Giulianini. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Giuseppina Adriana Buonamassa, Humanitas University, Rozzano, Italy
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.