How Artificial Intelligence Enhances Health Care

Abstract Artificial intelligence (AI) has been used for decades to mine clinical data, look for patterns, and produce results that can enable health care professionals to make decisions and be alerted rapidly to clinical problems that require attention. Refinements in algorithms, and massive increases in available data and speed of processing, have enabled real-time clinical decision making about diagnosis, prognosis, and therapy of individual patients. Moreover, large clinical datasets can be analyzed to identify and understand risk factors of disease for individuals or groups. A notable impact of AI is on the interpretation of images. For example, pathologic specimens can be examined for abnormalities, and their extent, chemical composition, and clinical significance ascertained. Thus far, the most studied application of AI is interpretation of radiologic images. AI’s performance rivals that of experienced radiologists and leads to more rapid and efficient generation of results. Based on these factors, guideline development is facilitated by generative algorithms. Based on guidelines and an individual’s clinical, environmental, and socioeconomic characteristics, patient-specific counseling can be generated. AI can influence publishing scientific information for good and ill. Large datasets can be combined, refined, and analyzed; statistical significance calculated; and all compared to the existing results and conclusions of others. However, generative AI can produce publications that embellish either inadequate or fraudulent data. Images can be manipulated or created with no basis in fact. Fortunately, just as AI can produce poor-quality or fake publications, AI programs are being developed to detect such nefarious activities.