AI code-writing assistants like OpenAI’s ChatGPT and GitHub Copilot have rapidly become part of programming education in higher education. This systematic literature review covers studies from 2015 to 2025 on integrating these AI coding tools into IT and computer science curricula, with a focus on: 1) potential challenges, 2) enhancements to learning, and 3) ethical considerations. Following PRISMA guidelines, we identified thirty-five relevant peer-reviewed articles and conference papers. The review finds that while AI assistants can provide on-demand help, personalized feedback, and new opportunities for student learning, they also introduce significant challenges related to academic integrity, student over-reliance, and the need to rethink assessment design. Key benefits reported include better debugging support, increased student engagement, and the potential for more personalized, self-paced learning. However, concerns such as plagiarism and weakened problem-solving skills are prominent. Ethical issues, ranging from fairness and access disparities to data privacy and the need for responsible-use policies, are discussed based on current literature and guidelines. The study concludes with practical recommendations for educators and institutions on harnessing AI code assistants to improve learning while mitigating risks, and it suggests directions for future research in this rapidly evolving domain.