A review of explainable artificial intelligence in smart manufacturing

Artificial Intelligence (AI) technologies have become essential in smart manufacturing, driving predictive capabilities and operational efficiency. However, the opacity of AI decision-making remains a critical barrier, as it limits interpretability and trust in high-stakes manufacturing environments. Explainable AI (XAI) addresses this challenge by making AI models more interpretable and trustworthy. Yet, due to the relative novelty of XAI, there are substantial challenges in implementation, a lack of standardised frameworks, and limited methods for quantitative evaluation. As a result, current applications of XAI in smart manufacturing remain under-developed, non-standardised, and fragmented. This review thus aims to provide a comprehensive exploration of the current landscape of XAI, highlighting recent advancements and critically examining its role in enhancing trust and transparency in smart manufacturing. Given the increasing reliance on AI for decision-making in complex manufacturing systems, a focused review of XAI is crucial for identifying pathways to more transparent and responsible AI-driven solutions. The paper also discusses key implementation challenges and outlines future research directions, with insights into how XAI could shape the future of smart manufacturing.