This paper addresses the hazards and benefits that large language models (LLMs), such as ChatGPT, bring to the instruction of Academic Writing for Computer Science students. On the one hand, advancements in artificial intelligence (AI) may decrease students’ motivation to study academic writing and tempt them to generate texts using chatbots. This reluctance to invest effort in practicing academic writing can question the learning outcomes in writing courses. On the other hand, AI can aid students in achieving greater autonomy by enabling them to check their writing independently. What is more, ChatGPT can reduce instructors’ workload by helping them produce teaching materials and answer the students’ simple questions that do not require expertise. We, therefore, conclude that LLMs are inevitable, and writing instructors should teach students to use AI responsibly to benefit from its powers while avoiding unadvisable consequences of the reckless use of AI.