The integration of Artificial Intelligence (AI) into the justice system has significantly improved processes like criminal risk assessment and the automation of judicial decisions. However, it also brings ethical and regulatory concerns that cannot be overlooked. This study, following the PRISMA 2020 methodology and the SPICE framework, reviews 19 selected studies from a total of 174 articles sourced from Scopus and ScienceDirect (2020–2025, Q1). The findings highlight that while AI enhances judicial efficiency, implementing it without clear regulations can undermine fairness, transparency, and accountability. Key risks include algorithmic bias, limited explainability, and inconsistencies in regulatory approaches across jurisdictions. To address these challenges, the study advocates for global regulatory harmonization and stronger measures to ensure AI-driven judicial decisions remain transparent, fair, and subject to human oversight.