Doctoral student's strategy use in GAI chatbot-assisted L2 writing: An activity theory perspective

Generative artificial intelligence (GAI) chatbots (e.g., ChatGPT) have significantly reshaped the landscape of writing, presenting both opportunities and challenges for L2 writers, especially at advanced academic levels. However, limited research has examined the strategies learners adopt when they proactively use GAI chatbots to support L2 writing. To fill the void, this qualitative case study explored seven doctoral students’ strategy use in GAI chatbot-assisted writing from an activity theory perspective. Data were collected from multiple sources, including chat logs between students and ChatGPT, semi-structured interviews, and student writing drafts. Data analysis showed that the doctoral students formed a writing community with ChatGPT, in which they used six strategies to boost the writing process and enhance writing quality. These strategies included using code-mixed prompts, employing writing conventions, adhering to research ethics, ensuring information security, seeking help from others, and assuming multiple roles with a human-centered approach. The findings offer valuable insights into the complex interplays between doctoral students and GAI chatbots in GAI chatbot-assisted L2 writing and shed useful light on the responsible use of GAI chatbots in doctoral writing.