Model selection is an extremely important area of statistics in which statisticians must first decide on a model selection criterion and then utilize that criterion in an effort to determine the “best” model for solving the problem at hand. While there are several choices when it comes to model selection techniques, we will focus on just one: Akaike Information Criterion (AIC). A light introduction to the likelihood paradigm is necessary in order to establish notation before going into the details of AIC as a likelihood-based method. We present AIC as a suitable method for model selection while highlighting the lack of an alternative that falls strictly under the likelihood paradigm. Model averaging is an additional utilization of AIC in which several candidate models can be used simultaneously in both inferential and prediction settings. Finally, we provide a comparison of model selection techniques in a prediction setting and introduce areas of further research.