Teaching Writing with GenAI: Teacher Insights on How to Adapt Learning and Assessment Practices

Generative artificial intelligence (GenAI) creates two connected challenges for English as a foreign language (EFL) writing assessment: teachers need credible evidence of learner authorship and learning, while teachers and learners need practical guidance for responsible use. We investigated how nine Vietnamese tertiary EFL teachers redesigned writing tasks and assessment and whether their pedagogical reasoning varied across career stages. We generated data through semi-structured interviews at two tertiary institutions and analyzed the accounts using reflexive thematic analysis. The analysis produced four themes: (1) movement from detection to authenticity-by-design, (2) Defensive, Adaptive, and Generative response patterns, (3) integrity, equity, and feasibility as design trade-offs, and (4) learner learning as learning to show, say, and cite. At the design stage, we expected career stage to help explain differences in teacher responses. The findings showed that career stage operated as one part of teachers’ professional histories, while course ecology, assessment stakes, class size, infrastructure, and institutional guidance often had the most immediate influence. The study contributes a teacher-response typology and a process model for designing assessment that makes learner judgment, source use, revision, and AI assistance assessable.