(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Ethical challenges and evolving strategies in the integration of artificial intelligence into clinical practice [1]

['Ellison B. Weiner', 'Department Of Biomedical Data Science', 'Geisel School Of Medicine', 'Dartmouth College', 'Hanover', 'New Hampshire', 'United States Of America', 'Irene Dankwa-Mullan', 'Department Of Health Policy', 'Management']

Date: 2025-05 

Artificial intelligence (AI) has rapidly transformed various sectors, including healthcare, where it holds the potential to transform clinical practice and improve patient outcomes. However, its integration into medical settings brings significant ethical challenges that need careful consideration. This paper examines the current state of AI in healthcare, focusing on five critical ethical concerns: justice and fairness, transparency, patient consent and confidentiality, accountability, and patient-centered and equitable care. These concerns are particularly pressing as AI systems can perpetuate or even exacerbate existing biases, often resulting from non-representative datasets and opaque model development processes. The paper explores how bias, lack of transparency, and challenges in maintaining patient trust can undermine the effectiveness and fairness of AI applications in healthcare. In addition, we review existing frameworks for the regulation and deployment of AI, identifying gaps that limit the widespread adoption of these systems in a just and equitable manner. Our analysis provides recommendations to address these ethical challenges, emphasizing the need for fairness in algorithm design, transparency in model decision-making, and patient-centered approaches to consent and data privacy. By highlighting the importance of continuous ethical scrutiny and collaboration between AI developers, clinicians, and ethicists, we outline pathways for achieving more responsible and inclusive AI implementation in healthcare. These strategies, if adopted, could enhance both the clinical value of AI and the trustworthiness of AI systems among patients and healthcare professionals, ensuring that these technologies serve all populations equitably.

Copyright: © 2025 Weiner et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Introduction and motivation

In the last decade, artificial intelligence (AI) has made remarkable progress, primarily due to advancements in deep neural networks, widely known as deep learning. The transformative impact of these methodologies in non-medical fields, such as robotics, autonomous driving, and speech understanding, has fueled interest in their application to medicine. As a result, AI models have been seriously considered for use in adjacent domains, such as biomedicine and healthcare.

To date, these promising technologies have led to the creation of sophisticated AI systems capable of performing critical clinical tasks, such as medical image interpretation at the level of expert physicians [1–4]. Some of these innovative AI technologies have been developed by our team at Dartmouth [5–12]. Until recently, creating AI systems to assist pathologists, radiologists, and other imaging professionals required laborious feature engineering—the manual design of algorithms to preprocess images, segment anatomic structures, detect features, and classify abnormalities. Developing these systems often took years, but recent advances in AI have merged feature engineering with deep learning from large sets of labeled or even unlabeled training data [13]. Deep learning approaches are highly adaptable to diverse imaging tasks. These AI systems have the potential to transform healthcare delivery due to their ability to analyze large datasets and recognize complex patterns, making patient care more efficient and accurate. This can reduce diagnostic errors and healthcare costs and, most importantly, improve patient outcomes.

Over the last decade, electronic health records have become the most comprehensive source of clinical information for biomedical research and decision-making due to widespread adoption. However, much of this information exists in free-text format within clinical notes and reports. The variability and ambiguity of unstructured free text create major obstacles to rapid extraction and reuse of clinical data. These records may also lack control and negative cases and often contain significant gaps, as seen when patients transfer between hospital systems, raising concerns about data completeness. Recent advancements in natural language processing (NLP), fueled by the exceptional performance of large language models across diverse tasks, have opened new avenues for AI in this domain [14–21]. These NLP methodologies can address the challenges of extracting information from unstructured data, enabling the development of informatics methods that unlock valuable insights for translational research and clinical care [22–24].

Undoubtedly, medicine is undergoing a rapid transformation driven by AI and machine learning, a trend that will continue at an unprecedented rate. As these changes unfold, the medical community faces pressing concerns regarding the ethical implementation of AI while legislation struggles to keep pace. AI ethics is a field focused on the responsible development, deployment, and use of AI within the constraints of current legal and ethical standards. The rapid pace of AI innovation demands an inclusive discussion among experts to ensure ethical use, prompting a surge of research in this area. No single group should dictate the solutions to these complex issues, nor should discussions become insular. Accordingly, this paper presents widely recognized challenges and proposed strategies for ensuring the ethical integration of AI into clinical practice. We will review the strengths of these recommendations and highlight areas that require further exploration.

[END]
---
[1] Url: https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000810

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/