The rapid development of artificial intelligence (AI) is fundamentally reshaping literary creation and writing instruction. This study proposes a practical framework for an AI-integrated science fiction creative writing course, centered on a "hierarchical technical architecture" and a "non-substitutive principle." By combining large language models with multimodal generative tools, this approach systematically addresses traditional pedagogical challenges like creative blocks and delayed feedback. Based on a course taught at Chongqing University, the course demonstrates how AI can support five key stages of the writing process: world-building, story outlining, scene description, draft revision, and short play adaptation. This method spans the entire process from text creation to visual expression, showcasing the potential and challenges of technology-empowered education. Additionally, the course incorporates targeted measures to address issues such as creative atrophy and ethical concerns that arise from over-reliance on AI.

artificial intelligence  /  science fiction  /  creative writing  /  pedagogical reform