The integration of artificial intelligence (AI) in higher education, encompassing both traditional AI systems (automatic or semi-automatic learning processes) and generative AI (such as large language models like ChatGPT), has transformed teaching, learning, and administration. While adaptive learning systems and generative AI tools enable personalized education and improved efficiency, they also raise ethical concerns, including privacy, bias, and over-reliance. However, there is a lack of guidelines to ensure transparency and equity in automated evaluation processes. Current studies focus on the pedagogical benefits of AI, but end up leaving aside important issues, such as algorithmic bias and the impact on academic integrity. This paper examines these challenges and the shifting roles of educators and students, emphasizing the dual impact of AI. It advocates for responsible adoption to balance benefits with human values and provides practical recommendations to guide stakeholders in navigating AI's transformative potential in education.