The next-generation polymer composite revolution: AI-driven multi-functional polymer composite for intelligent additive manufacturing

Polymer composites have involved a trade-off between exceptional specific stiffness and strength and complex, energy-intensive processes with limited functional applications. Intelligent additive manufacturing (AM) unlocks geometric and processing limitations that have been constrained by anisotropic behaviour and stringent performance requirements. Artificial intelligence (AI)- embedded AM can overcome those technical limitations, enabling precise programming and multi-functional structures. This review article examines machine learning accelerates the discovery of additively manufactured polymer composites through inverse design, predictive material-process-property relationships, and machine learning-assisted metamaterials. This article describes the AI-assisted control system, including intelligent monitoring systems, digital twin collaborative virtual reality, and AI-driven optimization, which ensures a reliable fabrication process for AI-driven multi-functional structures. Moreover, this article discusses multi-functional structures, including intelligent stimulus-responsive 4D transformable structures and self-sensing structures with structural health monitoring, which integrate capabilities such as energy harvesting, thermal management, biomedical and soft robotics, and wearables beyond load-bearing performance. These advances enable lightweight, efficient, and intelligent additive manufacturing systems for practical applications, ranging from personalised medical implants to next-generation engineering. Furthermore, critical challenges and future perspectives are identified to provide new insights into AI-driven multi-functional polymer composites for intelligent additive manufacturing. The machine learning-assisted polymer composite additive manufacturing can catalyse a paradigm shift from passive load-bearing structures to active intelligent systems with integrated self-sensing, self-adaptation, and self-healing capabilities, ultimately enabling full-scale functional programming from molecular design to macroscopic characterisation within closed-loop intelligent manufacturing frameworks.