Antimicrobial stewardship programs (ASPs) are essential in optimizing the use of antibiotics to address the global concern of antimicrobial resistance (AMR). Artificial intelligence (AI) and machine learning (ML) have emerged as promising tools for enhancing ASPs efficiency by improving antibiotic prescription accuracy, resistance prediction, and dosage optimization. This systematic review evaluated the application of AI-driven ASPs, focusing on their methodologies, outcomes, and challenges. We searched all of the databases in PubMed, Scopus, Web of Science, and Embase using keywords related to "AI" and "antibiotic." We only included studies that used AI and ML algorithms in ASPs, with the main criteria being empirical antibiotic selection, dose adjustment, and ASP adherence. There were no limits on time, setting, or language. Two authors independently screened studies for inclusion and assessed their risk of bias using the Newcastle Ottawa Scale (NOS) Assessment tool for observational studies. Implementation studies underscored AI's potential for improving antimicrobial stewardship programs. Two studies showed that logistic regression, boosted-tree models, and gradient-boosting machines could effectively describe the difference between patients who needed to change their antibiotic regimen and those who did not. Twenty-four studies have confirmed the role of machine learning in optimizing empirical antibiotic selection, predicting resistance, and enhancing therapy appropriateness, all of which have the potential to reduce mortality rates. Additionally, machine learning algorithms showed promise in optimizing antibiotic dosing, particularly for vancomycin. This systematic review aimed to highlight various AI models, their applications in ASPs, and the resulting impact on healthcare outcomes. Machine learning and AI models effectively enhance antibiotic stewardship by optimizing patient interventions, empirical antibiotic selection, resistance prediction, and dosing. However, it subtly draws attention to the differences between high-income countries (HICs) and low- and middle-income countries (LMICs), highlighting the structural difficulties that LMICs confront while simultaneously highlighting the progress made in HICs.