The rapid digitalization of the energy sector has accelerated the use of Artificial Intelligence (AI) for forecasting, control, fault detection, and decision support. However, increasing system complexity, safety requirements and uncertainty have exposed the limits of fully autonomous AI in real-world energy environments. Human-in-the-Loop Artificial Intelligence (HITL-AI) has emerged as a promising direction that integrates human expertise into AI development and operation, offering better transparency, adaptability, and reliability, which are key principles of the industry 5.0 vision. This review provides a systematic synthesis of HITL-AI technologies applied across the energy domain, including active learning, interactive machine learning, explainable AI, reinforcement learning, curriculum learning and recent agentic AI frameworks. Motivations behind HITL-AI adoption and its main application areas were analyzed. The review also identifies operational drivers shaping the future of human–AI collaboration in energy systems. Finally, limitations and strategic research directions needed to enable safe, trustworthy, and HITL-AI were outlined for next-generation energy infrastructures. This work laid a foundation knowledge from the past decade and offers a comprehensive roadmap for advancing HITL-AI in support of resilient and sustainable energy systems aligned with Industry 5.0.
Nguyen et al. (Sun,) studied this question.