Do Multilingual Language Models Capture Differing Moral Norms?
This addresses potential cultural biases and harmful outcomes in multilingual AI systems, particularly for low-resource languages, but is incremental as it builds on existing work in model evaluation.
The study investigated whether multilingual language models capture differing moral norms across languages, finding that models like XLM-R do capture moral norms, sometimes with higher human agreement than monolingual models, though the extent of differences between languages remains unclear.
Massively multilingual sentence representations are trained on large corpora of uncurated data, with a very imbalanced proportion of languages included in the training. This may cause the models to grasp cultural values including moral judgments from the high-resource languages and impose them on the low-resource languages. The lack of data in certain languages can also lead to developing random and thus potentially harmful beliefs. Both these issues can negatively influence zero-shot cross-lingual model transfer and potentially lead to harmful outcomes. Therefore, we aim to (1) detect and quantify these issues by comparing different models in different languages, (2) develop methods for improving undesirable properties of the models. Our initial experiments using the multilingual model XLM-R show that indeed multilingual LMs capture moral norms, even with potentially higher human-agreement than monolingual ones. However, it is not yet clear to what extent these moral norms differ between languages.