<p>In the midst of uncertainties and wavering commitments over a fully electric future, hybrid electric vehicles (HEVs) have gained increasing attention as a sustainable solution to reduce fuel consumption and emissions in the transportation sector. At the core of their performance is the Energy Management Strategy (EMS), which governs the efficient distribution of power among internal combustion engines (ICE), electric motors (EM), and battery systems. Conventional EMS approaches, such as rule-based and optimisation-based strategies, often struggle to adapt to the variability of real-world driving conditions. In response, Artificial Intelligence (AI)-based techniques have emerged as powerful alternatives, offering data-driven, adaptive, and enhanced energy distribution control. This systematic review analyses AI-based EMS for HEVs using studies from 2015 to 2025 sourced from Google Scholar, Springer, and IEEE Xplore. A PRISMA-compliant screening process yielded 32 studies for synthesis. The review focuses on fuzzy logic (FL), neural networks (NN), reinforcement learning (RL), and hybrid models combining heuristic and learning-based methods. Across the reviewed literature, FL achieves moderate fuel efficiency improvements with the highest interpretability. NN approaches report high fuel reduction with strong generalisation capability while RL methods demonstrate superior adaptability to previously unseen driving conditions, though at higher computational cost. Future research directions are discussed including Explainable AI (XAI) and Human-in-the-Loop (HITL) frameworks aligned with Industry 5.0 principles, aimed at developing more transparent, user-centric, and deployable EMS for next-generation HEVs.</p>