This column introduces four articles focusing on GAI-empowered foreign language teaching, covering human-AI collaborative design of continuation writing in intensive French reading courses, translingual practices of GAI chatbots in English classrooms, value-added effects of GAI feedback on English writing, and rational understanding of educational digitalization, aiming to provide frontline teachers with theoretical references and practical pathways for technology-integrated instruction while emphasizing the critical role of teacher professional guidance in stimulating learner agency.

1　Introduction

This section introduces the concept of continuation writing tasks and their learning mechanisms, noting that while widely applied in English teaching, research on their implementation in zero-beginner foreign language intensive reading courses such as French remains insufficient. It then identifies two major challenges in implementing such tasks in intensive reading courses—design appropriateness and feedback effectiveness—and clarifies new pathways for addressing these issues through large language model tools. Finally, it specifies that this study takes second-year French intensive reading courses as its context to explore task design, implementation, and optimization pathways from a human-AI collaborative perspective, focusing on two core issues: AI-assisted task design and feedback optimization.

2　Literature Review

This section reviews the theoretical foundations of continuation writing tasks, empirical evidence, and applications of AI tools in foreign language education. Continuation writing centers on alignment effects, emphasizing that learners achieve application and transfer of linguistic knowledge through creative imitation via deep cognitive interaction with preceding texts. Numerous empirical studies have validated its facilitative effects on improving lexical diversity, syntactic complexity, and form-meaning mapping abilities, with the key lying in learners' deep processing of discourse clues and communicative intentions. The task's applications in teaching and assessment differ: teaching should highlight target words and encourage reuse, while assessment requires using one's own language. When applied in intensive reading courses, the focus should be on contextualized reuse of learned knowledge, but difficulties in compiling preceding texts constrain its practice. AI tools provide technical support for foreign language education through context generation, multimodal resource creation, and immediate feedback, capable of generating comprehensible input materials and providing multidimensional writing feedback. However, their effective application depends on teachers' human-AI interactive negotiation abilities and critical technology literacy, requiring clear task boundaries for teachers, AI, and human-AI collaboration to form a bidirectional-stimulation human-AI collaborative co-education model.

3　Instructional Case Design and Implementation

This section presents the design and implementation process of a human-AI collaborative continuation writing task in second-year French intensive reading courses. With preceding text generation as the key, researchers adopted an indirect strategy after multiple attempts: first having AI generate complete stories, then rewriting to preserve only the beginning and development sections, extracting approximately 40–50 words as preceding materials to prompt students to naturally reuse seven emotion-expression target vocabulary items when continuing the climax, conclusion, and ending. The implementation was conducted extracurricularly with six students on a small scale, completed under offline supervision, allowing reference materials but prohibiting AI-assisted writing. Evaluation employed a dual mechanism of AI preliminary assessment and teacher final assessment: AI focused on communicative purpose, emotional expression coherence, and situational appropriateness, while teachers attended to grammatical and lexical accuracy and target vocabulary reuse. Documents were compiled and individual face-to-face feedback provided. Data collection encompassed teacher-AI interaction logs, 12 student continuation compositions, AI feedback documents, teacher annotations, and field notes.

4　Research Findings and Discussion

This section examines the design effects of intensive-reading-embedded continuation writing tasks and application limitations of AI tool feedback. Regarding task design, target vocabulary transfer rates in student continuations were extremely low, but performance in plot integrity, logical coherence, and reuse of prior knowledge was satisfactory, revealing differences between intensive-reading-embedded tasks and learning-facilitation tasks: the former focuses on learner contexts, emphasizing rolling accumulation and holistic language competence development, while the latter triggers specific structures through closed contexts. Regarding task evaluation, students recognized AI feedback as pertinent but considered its examples beyond current proficiency levels and lacking specificity; AI feedback relies on cross-sectional data and cannot grasp student developmental trajectories or instructional backgrounds, whereas teacher feedback possesses dynamic diagnostic advantages. The optimal solution should establish a human-AI collaborative mechanism, with AI assuming immediate surface-level feedback and teachers implementing deep diagnostic evaluation, balancing efficiency and professionalism.

5　Conclusion and Implications

This section explores the design logic and evaluation mechanisms of intensive-reading-embedded continuation writing tasks from a human-AI collaborative perspective. Regarding task design, AI can rapidly generate appropriate preceding texts, but open tasks cannot guarantee targeted reuse; if focusing on specific linguistic phenomena is needed, texts highlighting target language points can be generated, while if enhancing creativity is desired, implicit context texts can be generated. Regarding post-writing feedback, AI feedback is immediate and comprehensive but insufficiently specific, unable to capture student dynamic development; teacher-based deep diagnosis from longitudinal observation remains irreplaceable. Human-AI collaboration combines AI efficiency with teacher professional judgment, instead highlighting the importance of the teacher role. Future directions may encourage students to establish personal learning portfolios, interact regularly with AI, share progress trajectories, and achieve cyclic iteration and continuous improvement.