ETHICA-MT: Introducing a Framework and Dataset for Studying Ethical Orientations in LLM-based Machine Translation

ACL ARR 2026 January Submission4335 Authors

Keywords: Ethics in Machine Translation

Abstract: Translation is a fundamentally value-laden process that requires the translator to make decisions and judgments that have ethical implications. However, modern machine translation (MT) systems, including large language models, have not been systematically examined for their default ethical tendencies or their capabilities to accommodate specified ethical frameworks and priorities in conflicted translation situations. To address this gap, we present ETHICA-MT, a framework for examining ethical reasoning in MT systems. Drawing on established principles from the translation ethics literature, we formalize a conceptual framework and construct a multilingual benchmark ETHICA-MT BENCH that covers six languages and highlights ethical conflicts arising from competing ethical approaches in different translation scenarios. Our empirical study shows that current systems predominantly default to an ethical stance that favors `faithful representation' to the source text and highlights the challenges of automatically and manually judging ethical stances of the models.

Paper Type: Long

Research Area: Ethics, Bias, and Fairness

Research Area Keywords: Machine Translation, Ethics, ethical dilemma, faithful representation

Contribution Types: Model analysis & interpretability, Position papers

Languages Studied: English, Hindi, Bengali, Arabic, French, Hebrew

Submission Number: 4335