This systematic literature review (SLR) examines the limitations of Artificial Intelligence (AI) writing assistant tools and their implications for education. The study identifies six critical shortcomings: (1) lack of academic rigor, (2) insufficient knowledge base, (3) inability to synthesize complex ideas, (4) absence of human cognitive abilities, (5) limited adaptability, and (6) questionable originality. The review highlights the challenges of engaging students with diverse emotional and cognitive characteristics, supported by statistical insights into emotional and cognitive engagement. It explores the ethical considerations and potential impacts of AI technologies on academic integrity and credibility. The findings also encompass a comparative analysis of AI models like Gemini and ChatGPT, their applications in various educational contexts, and their effectiveness in tasks such as feedback mechanisms and assessment techniques. Findings underscore the tools' potential to enhance engagement and comprehension while emphasizing the need for balanced integration with human input to preserve academic credibility. This research highlights AI's transformative yet limited role in education, advocating for responsible use and continuous evaluation.