This paper presents a themed theoretical review on the benefits and drawbacks of implementing artificial intelligence and how to benefit from it to improve tax compliance approach and concentrates on Twenty-five recent previously conducted studies on this theme. Reading the various works, one understands that though AI could improve compliance, paper work and revenue collection, risks on ethics and privacy, and bias in the use of algorithms to make managerial decisions persist. It follows from these results that policy makers are faced with a moral dilemma of how to use advancements in technology without infringing on the rights of the taxpayers. Further studies should be conducted on the development of not only frameworks focusing on the efficiency of AI systems and its effectiveness within tax practices, but that also touch upon and address social challenges, especially among the vulnerable segments of society. In addition, it is equally important that the future of policy formulation takes up an interdisciplinary approach to ensure that code development does not go against fairness. Therefore, a preventive orientation is required to maximize advantages of AI while minimizing its risks.