The rapid advancement of large language models has led to an increase in the use of AI-generated text that is increasingly difficult to distinguish from human writing. Detecting AI-generated text has become crucial for maintaining the reliability of online content, combating misinformation, and preserving trust. This paper presents a comprehensive review of methods for AI-generated text detection. We review various detection techniques, systematically analyzing feature-based approaches, deep learning methods, unsupervised techniques, adversarial robust models, and ensemble frameworks. We critically evaluate the strengths and limitations of each methodology type. In addition, we survey the datasets and evaluation metrics commonly used to benchmark detection performance, and we discuss practical applications of AIgenerated text detection in domains such as education, social media, and journalism. The survey incorporates references to recent research developments and provides insights into how detection methods cope with increasingly sophisticated text generators. We conclude by summarizing current challenges – including the ongoing “arms race” between text generators and detectors – and highlighting promising directions for future research.