Artificial intelligence (AI) is reshaping algorithmic trading by enabling improved predictive power, smarter, more adaptive decisionmaking, and enhanced risk management. This work examines key AIdriven approaches, including Deep Learning (DL), Reinforcement Learning (RL), Sentiment Analysis (NLP), and Hybrid Models (HM), highlighting their applications, strengths, and challenges. It also explores the role of essential data sources, such as market data, event-driven data, fundamental indicators, and alternative data, which fuel AI-based trading strategies. Despite its advantages, AI still faces critical challenges related to data quality, computational complexity, and regulatory compliance. The paper concludes with future research directions, including Explainable AI (XAI), AutoML, and AI-Blockchain integration, which aim to overcome current limitations and support the development of more transparent, efficient, and robust intelligent trading systems.