Abstract: X-ray computed tomography (CT) is widely used in clinical practice, and image reconstruction algorithms play a pivotal role in determining image quality. With the increasing diversity of clinical imaging demands and concerns regarding radiation exposure, conventional reconstruction techniques, such as filtered back-projection and iterative reconstruction, have become insufficient for meeting the requirements of low-dose CT imaging and diversified diagnostic scenarios. In recent years, artificial-intelligence (AI) based reconstruction algorithms have been increasingly incorporated into clinical workflows, leading to substantial improvements in image quality and in radiation dose efficiency. In this review, first, the major technical bottlenecks associated with traditional CT reconstruction methods are outlined and the requirement to adopt AI-based reconstruction techniques in clinical practice is highlighted. Then, a systematic overview of the principal categories of AI-based CT reconstruction algorithms is provided, commonly used approaches are summarized, and their performance in clinical applications is described. Finally, the current limitations of AI-based CT reconstruction algorithms are discussed and potential strategies to advance their integration into clinical CT reconstruction are proposed.