With the growing frequency, scale, and sophistication of cyber threats, digital forensics has transitioned from largely manual investigative processes to intelligent, automated, and data-driven frameworks. Artificial Intelligence (AI) and Machine Learning (ML) have emerged as powerful tools for enhancing the detection, classification, and response to cyber-attacks, enabling near real-time analysis of vast and complex digital evidence. This review paper presents a comprehensive synthesis of recent research and practical advancements in the application of AI within digital forensics. It focuses on attack detection models, forensic analytical frameworks, performance strengths, inherent limitations, and emerging trends shaping the field. Supervised, unsupervised, and deep learning approaches are examined in detail, alongside hybrid and ensemble systems that combine multiple techniques to improve accuracy and resilience. Additionally, the study highlights key challenges in real-world deployment, including data quality, model interpretability, scalability, and ethical considerations, and discusses potential future directions for advancing AI-driven digital forensic investigations.