Artificial intelligence in healthcare: past, present and future

This survey examines the role of artificial intelligence, which aims to mimic human cognitive functions, in bringing a paradigm shift to healthcare powered by increasing availability of healthcare data and rapid progress in analytics techniques. The authors review how AI can be applied to both structured and unstructured healthcare data. Popular techniques include machine learning methods for structured data, such as classical support vector machines and neural networks, as well as modern deep learning, alongside natural language processing for unstructured data. They identify cancer, neurology, and cardiology as major disease areas where AI tools are used.

The paper then reviews AI applications in stroke in more detail, across three major areas: early detection and diagnosis, treatment, and outcome prediction and prognosis evaluation. It concludes with discussion of pioneer AI systems such as IBM Watson and the hurdles standing in the way of real-life deployment of AI. As a broad, accessible survey, it became a widely cited reference framing the landscape and challenges of AI in medicine.