Explainable artificial intelligence for laser powder bed fusion alloys: breaking the formability–performance trade-off via collaborative control

Laser powder bed fusion (LPBF) enables the fabrication of metallic components with complex geometries and high performance. However, achieving coordinated control of formability including defect suppression and cracking resistance together with mechanical properties such as strength, ductility, and fatigue life remains a critical challenge for LPBF alloys. Although artificial intelligence (AI) has been widely adopted for process optimization, defect prediction, and performance evaluation, conventional AI models often lack physical metallurgical interpretability, leading to limited generalization and unreliable extrapolation beyond training domains. This review focuses on the role of explainable artificial intelligence (XAI) in addressing the formability–performance trade-off in LPBF alloys. We first introduce four knowledge domains that establish physical interpretability in metallic materials: physical metallurgy principles, cross-scale numerical information, microstructure image analysis, and causal physical mechanisms. Subsequently, we systematically examine XAI applications across representative LPBF alloy systems. In performance-oriented alloys such as steels and titanium alloys, XAI facilitates mechanism–aware control of non-equilibrium phase transformations, microstructural anisotropy, and defect-sensitive fatigue behavior. In formability-oriented alloys including aluminum alloys and nickel-based superalloys, XAI supports cracking susceptibility prediction, process–window identification, and printable alloy design. Despite these advances, most current studies remain task-specific and single-objective, with limited integration of formability and performance optimization. To overcome this limitation, we propose a unified multi-objective optimization framework incorporating physically constrained modeling, standardized multimodal databases, and causal reasoning. This review provides guidance for the intelligent, reliable, and collaborative design of LPBF alloys with both high formability and excellent mechanical performance.