From Compliance to Critical Literacy: Teaching GenAI Translation Ethics through Synthetic Scenarios

Abstract

The rapid integration of Generative AI (GenAI) into the translation industry has outpaced the development of ethical pedagogical frameworks, leaving students ill-equipped to navigate complex human-machine collaborations. This study addresses this gap by proposing a Situated Learning approach to the teaching of GenAI ethics. Grounded in the 2025 TAC Guidelines for the Chinese translation industry, the study evaluates an instructional design that utilises synthetic, scenario-based learning materials.

As part of a Teaching Development Grant project, 72 translation students (BA and MA) engaged with a series of AI-generated ethical dilemmas focusing on transparency, risk management, and intellectual property. The intervention employed instructional scaffolding through immediate feedback loops (Moodle Lessons) followed by reflective inquiry. Results indicate that while the scaffolding effectively solidified surface knowledge of normative constraints, students struggled with ‘grey area’ decision-making, particularly regarding fan translation and copyright. Qualitative analysis shows students’ understanding evolving from passive compliance to active Critical AI Literacy (CAIL) through the intervention. The study concludes that synthetic scenarios are a viable pedagogical tool for fostering moral imagination and supporting the teaching and learning of GenAI Translation Ethics.