Investigating Markers and Drivers of Gender Bias in Machine Translations

Implicit gender bias in Large Language Models (LLMs) is a well-documented problem that needs to be better understood in order to be addressed effectively.Implications of gender introduced into automatic translations can perpetuate real-world biases in Software Engineering and other domains.However, some LLMs use heuristics or post-processing to mask such bias, which makes investigation more difficult.Here, we examine bias in language models via back-translation, using the DeepL online translation service to investigate the bias evinced when repeatedly translating a set of 56 Software Engineering tasks used in a previous study.Each statement starts with 'she', and is translated first into a 'genderless' intermediate language then back into English; we then examine pronounchoice in the back-translated texts.We believe this approach provides a useful alternative to large-scale surveys in mapping biases.We expand prior research in the following ways: (1) by comparing results across five intermediate languages, namely Finnish, Indonesian, Estonian, Turkish and Hungarian; (2) by proposing a novel metric for assessing the variation in gender implied in repeated translations of the same phrase, avoiding the over-interpretation of individual pronouns, apparent in earlier work; (3) by investigating sentence features that drive bias; (4) and by comparing results from three time-lapsed datasets to establish the reproducibility of the approach.We found that some languages display similar patterns of pronoun use, falling into three loose groups, but that patterns vary between groups; this underlines the need to work with multiple languages.We also identify the main verb appearing in a sentence as a likely significant driver of implied gender in the translations.Moreover, we see a good level of replicability in the results, and establish that our variation metric proves robust despite an obvious change in the behaviour of the DeepL translation API during the course of the study.These results show that the back-translation method can provide further insights into bias in language models.