GenAI-mediated translation: Toward developing an actionable ethical guideline for translators

Abstract Actionable ethical guidelines are crucial for enabling translators to engage with GenAI responsibly. However, the available AI ethical guidelines in translation neither adequately discuss the ethical issues that may arise nor provide sufficiently actionable guidance for practicing translators. This study thus attempts to develop a practical ethical guideline for GenAI-mediated translation, in response to recent calls to operationalize AI ethics. This objective was pursued through three steps. First, we reviewed the relevant literature to derive ethical principles for the use of GenAI in translation. Second, we conducted moderated focus groups with a broad range of translation stakeholders to identify concrete practices through which each principle could be put into practice. Third, we employed a Delphi consensus-building process with panelists drawn from the focus group participants to prioritize the principles and practices for inclusion in the guideline. The outcome is a proposed ethical guideline consisting of a cascade of 9 ethical principles and 59 recommended practices, all ranked by relative importance. While providing a resource that translators can immediately adopt, this study also calls for further research to explore potential barriers to its implementation and to localize the guideline for specific translation domains and countries, where appropriate.