The rapid growth of e-commerce has made it an attractive target for cybercriminals, jeopardizing consumer trust and the financial stability of businesses. In response, organizations are increasingly turning to AI and machine learning to enhance data security. This literature review examines both the theoretical and practical dimensions of AI-driven e-commerce security, focusing on frameworks such as threat modeling and the zero trust security model. It evaluates the application of supervised, unsupervised, and adversarial learning techniques, highlighting their effectiveness in detecting both known and emerging threats. Key applications of AI and machine learning include real-time threat detection, user authentication, fraud prevention, and incident response. The review also addresses challenges such as data privacy concerns, adversarial attacks, scalability, and interpretability. It concludes by underscoring the necessity for a balanced approach that integrates theoretical insights, practical strategies, and ongoing innovation to establish a robust security infrastructure for e-commerce. By harnessing these technologies, businesses can better safeguard customer data, sustain trust, and protect their platforms against evolving cyber threats.