In this manuscript, Yao et al. review the applicability of machine learning (ML) models in predicting outcomes or sequelae from patients with early clinical presentation of mild traumatic brain injury (mTBI). Of the 1235 initially identified papers (published until 10 March 2023), 10 met the inclusion criteria for this study, representing data from 127,929 patients. The limited number of included studies reveals the low consensus and consistency in assessing and reporting mTBI recovery or complications. Even among the 10 studies selected for further analysis, issues related to data quality, risk of bias, and minimal overlap between studies, prevented generalizations and limited the conclusions drawn. Nevertheless, the two most frequently used ML modeling techniques (support vector machine - SVM; and Artificial neural network - NN) showed promising results, outperforming traditional statistic methods in the two studies where such comparisons were made. Overall, this work reveals the potential of ML for prognostic and managing mTBI recovery, while calling for more studies and higher consistency in assessing, reporting, and data availability regarding mTBI patient’s trajectory.