Artificial intelligence (AI) and medical informatics research fields have considerable overlap, with technologies supporting different health issues in different contexts. In this work, we aimed to map out and understand the contributions of AI in medical informatics over time. To that, we applied bibliometric analysis with scientific literature since the 1970s. The production of papers exponentially increased over time, and we found periods with similar characteristics of the content. We also identified different clusters of technologies and applications varying according to the periods and related keywords. We hypothesized some future directions for the use of AI in medical informatics.