<![CDATA[Video-based artificial intelligence has emerged as a rapidly growing field, driven by advancements in deep learning and the increasing demand for automation across sectors. This study aims to summarize the trends, applications, and major challenges in the implementation of video-based AI using a PRISMA-based systematic literature review approach. The data synthesized from 17 selected articles indicates that deep learning models such as CNNs, LSTMs, and hybrid architectures have been successfully employed for various tasks including anomaly detection, deepfake classification, long-range surveillance, video compression, and video-based educational assessment. Applications span across security, healthcare, education, and entertainment, with notable improvements in efficiency and accuracy. However, challenges remain concerning privacy, algorithmic bias, and the gap between technological progress and regulatory readiness. Hardware demands and variability in model performance also pose limitations. These findings underscore the importance of interdisciplinary approaches to foster responsible and sustainable innovation in video-based AI. The review offers a comprehensive overview that may serve as a foundation for future research directions and technological development.]]>
Copyrights © 2025
