Artificial Intelligence in Medicine: Potential Benefits and Risks

A recent field of science and technology is artificial intelligence (AI). It already significantly impacts various human activities at all societal scales, ranging from individuals to social organizations, corporations, and countries. Almost every industrial, economic, and societal sector is experiencing rapid growth in artificial intelligence, from information technologies to commerce, manufacturing, space, remote sensing, security, defense, transport, and vehicles, and, since the turn of the century, medicine and health care.1 Deep learning tools and AI are expected to have an impact on healthcare on three different levels: for clinicians, primarily through quick, accurate image interpretation; for health systems, through improved workflow and the potential to lower medical errors; and for patients, through the ability to process their data to improve health.2 Soon, almost every clinician will be utilizing AI technologies. Deep neural networks (DNNs), which can read medical scans, pathology slides, skin lesions, retinal pictures, electrocardiograms, endoscopy, and many more domains, were primarily used for pattern recognition in this. Radiology is one area where artificial intelligence has received much attention. Over two billion chest X-rays are performed globally each year, making them the most used type of diagnostic scan. One study evaluated the performance of one deep neural network and four radiologists in identifying pneumonia in over 112,000 tagged frontal chest X-ray images. The results showed that the algorithm could outperform the radiologists.