Mammography imaging is one of the most demanding imaging modalities from the point of view of the balance between image quality (the visibility of small size and/or low contrast structures) and dose (screening of many asymptomatic people). Therefore, since the introduction of the first dedicated mammographic units, many efforts have been directed to seek the best possible image quality while minimizing patient dose. The performance of automatic exposure control (AEC) is the manifestation of this demand. The theory of AEC includes exposure detection and optimization and also involves some accomplished methodology. This review presents the development and present situation of spectrum optimization, detector evolution, and the way how to accomplish and evaluate AEC methods.