Artificial Intelligence in Healthcare and Medicine: From Machine Learning to Autonomous Agents | RISE Research

This review paper examines the growing role of artificial intelligence in healthcare and medicine, tracing its evolution from foundational machine learning and deep learning techniques through natural language processing, large language models, and emerging autonomous AI agents. The paper covers how supervised, unsupervised, and reinforcement learning are applied to clinical tasks, with particular attention to deep learning's success in medical imaging (detecting abnormalities like early breast cancer, pneumonia, and fractures in X-rays, CT, and MRI scans at accuracy sometimes comparable to trained specialists), predictive patient monitoring for conditions like sepsis and cardiac arrest, and precision medicine using genetic and biomarker data. It then extends into how NLP and large language models process the substantial share of medical data that exists as unstructured text (physician notes, discharge summaries, pathology reports), and how emerging agentic AI systems can support more complex, multi-step clinical workflows and administrative tasks. Alongside these applications, the paper systematically addresses key limitations and risks: data quality and representativeness issues that can cause AI systems to reinforce existing healthcare disparities for underrepresented populations, the 'black box' interpretability problem in deep learning models that complicates clinical trust and adoption, data security and cybersecurity concerns, and the reduced human interaction that heavier AI reliance can introduce into patient care. The paper concludes that AI in healthcare is best understood as a tool designed to assist, not replace, doctors and nurses, augmenting clinical decision-making and improving the quality and efficiency of care delivery, with its role expected to keep expanding as the underlying technology matures.