The dramatic growth of digital data, along with the rising implementation of cloud computing and IoT devices, has posed substantial difficulties in digital forensics investigations. Standard forensic techniques struggle to face data distribution, multi-tenancy, device diversity, and the extensive quantity of network traffic, complicating the collection, analysis, and interpretation of digital evidence highly challenging. This study delves into the integration of artificial intelligence, particularly Machine Learning and Deep Learning, to streamline and optimize forensic procedures in three primary fields, Network Forensics, Cloud Forensics, and IoT Forensics. Through a comprehensive academic survey, the research pinpoints domain-specific challenges, analyzes intelligent approaches, and reviews the role of different datasets. Findings propose that artificial intelligence methods considerably improve evidence identification, anomaly detection, and forensic automation, as a result, lowering manual effort and improving investigation accuracy. However, challenges such as standardization, data privacy, and adversarial AI threats remain major obstacles to the complete implementation of AI. The paper finalizes by presenting future research directions, underlining the necessity of robust AI models, forensic standardization, and scalable analytics to improve the reliability and effectiveness of AI-driven digital forensics.