This research examines the drawbacks associated with ChatGPT, a smart writing assistant chatbot, within the context of student writing. Employing qualitative methods, notably in-depth interviews with 30 participants, the study meticulously analyzes the challenges and constraints entailed in relying on ChatGPT for writing assistance. Centered around five key themes — understanding, dependency, feedback, writing style, and plagiarism — the findings clarify that students struggle with comprehending ChatGPT's output, potentially leading to confusion and misinterpretation. The feedback provided by the AI writing assistant emerges as generic and lacking specificity, constraining its efficacy in guiding students' writing processes. Participants report an obvious shift towards more predictable and impersonal writing styles attributed to their usage of ChatGPT—furthermore, apprehensions surface concerning the unintentional risk of plagiarism when leveraging the tool. The implications underscore educators' need to acknowledge the limitations of AI writing assistants, offering targeted support to help students navigate these challenges while fostering critical thinking, creativity, and independent writing skills. The findings promote ongoing research and improvement in AI writing assistants to improve understanding, elevate feedback quality, and address identified limitations. By mitigating these drawbacks, educators, students, and developers can optimize the utility of AI writing assistants, ensuring they serve as valuable tools for supporting writing development while upholding critical thinking, creativity, and ethical writing practices. This research not only reveals the drawbacks of ChatGPT as a writing assistant but also advocates for continual research and development in the field of AI writing assistance.