With the increasing complexity of the electricity network, optimizing short-term planning for unit commitment (UC) demands more time and resources. This has prompted many researchers in the field to explore alternative methods, serving as auxiliary accelerators or even replacements for classical Mixed Integer Programming (MIP) algorithms. The core idea behind these alternatives is to leverage the fact that the UC problem is typically solved daily by Independent System Operators (ISOs) with minor changes, resulting in a substantial amount of available data. This characteristic makes the problem particularly suitable for Artificial Intelligence (AI) techniques, such as Machine Learning (ML). This paper aims to delineate the state of the art of AI applied to the UC problem, categorizing articles based on the different AI techniques employed and the characteristics and complexity of the UC problem addressed. Additionally, it seeks to compare the outcomes achieved. As of the author’s knowledge, this systematization task has not yet been accomplished in this emerging research area.