Artificial intelligence (AI) is poised to broadly reshape medicine, potentially improving the experiences of both clinicians and patients. We discuss key findings from a 2-year weekly effort to track and share key developments in medical AI. We cover prospective studies and advances in medical image analysis, which have reduced the gap between research and deployment. We also address several promising avenues for novel medical AI research, including non-image data sources, unconventional problem formulations and human–AI collaboration. Finally, we consider serious technical and ethical challenges in issues spanning from data scarcity to racial bias. As these challenges are addressed, AI’s potential may be realized, making healthcare more accurate, efficient and accessible for patients worldwide.
In the years ahead, AI is poised to broadly reshape medicine. Just a few years since the first landmark demonstrations of medical AI algorithms that are able to detect disease from medical images at the level of experts1,2,3,4, the landscape of medical AI has matured considerably. Today, the deployment of medical AI systems in routine clinical care presents an important yet largely unfulfilled opportunity, as the medical AI community navigates the complex ethical, technical and human-centered challenges required for safe and effective translation.
In this review, we summarize major advances and highlight overarching trends, providing a concise overview of the state of medical AI. Our review is informed by our efforts over the past 2 years, during which we tracked and shared recent developments in medical AI on a weekly basis (https://doctorpenguin.com). First, we summarize recent progress, highlighting studies that have rigorously demonstrated the utility of medical AI systems. Second, we examine promising avenues for medical AI research in the form of novel data sources and discuss collaboration setups between AI and humans, which are more likely to reflect real medical practice than typical study designs that pit AI against humans. Finally, we discuss major challenges facing the field, including the technological limitations of AI as it stands and ethical concerns about regulating AI systems, holding people accountable when AI error occurs, respecting patient privacy and consent in data collection and safeguarding against the reinforcement of inequities (Fig. 1).
To continue reading this article, click here.