Trust Me, I’m AI: A User-Centric Review of Human–Machine Interaction and Teaming With Explainable AI

As artificial intelligence (AI) systems increasingly permeate real-world decision-making scenarios, the demand for user-centered Explainable AI (XAI) becomes critical. This systematic review investigates the interdisciplinary landscape of user studies in XAI, focusing on the interplay between explanation design, trust formation, cognitive load, and human-AI teaming. Drawing from literature across machine learning, psychology, and human factors research, we analyze 31 empirical studies according to three guiding questions: 1) how XAI design affects user acceptance and understanding; 2) which psychological and cognitive factors influence trust development; and 3) how AI explainability impacts teamwork, decision quality, and situational awareness in collaborative human-AI settings. We provide a critical synthesis of experimental designs, participant demographics, evaluation metrics, and replicability standards, revealing common methodological shortcomings such as age and gender imbalances, over-reliance on student samples, and inconsistent reporting practices. Based on our findings, we propose the “Design–Trust–Teaming” (DTT) framework to guide future research and evaluation of user-centered XAI systems. This review offers actionable insights for designing adaptive, inclusive, and trustworthy AI systems that facilitate meaningful collaboration between humans and intelligent agents.