This paper tries to quantify the ethical dilemma of using culturally toxic training data to improve the performance of AI tools for ultra low-resource languages such as Indigenous languages. Our case study explores the use of Bible data which is both a commonly available source of training pairs for translators of Indigenous languages and a text which has a trail of physical and cultural violence for many Indigenous communities. In the context of fine-tuning a WMT19 German-to-English model into a Guarani Mbya-to-English translator, we first show, with two commonly-used Machine Translation metrics, that using only Bible data is not enough to create successful translators for everyday sentences gathered from a dictionary. Indeed, even fine-tuning with only 3,000 pairs of data from the dictionary produces significant increases in accuracy compared to Bible-only models. We then show that simultaneously fine-tuning with dictionary and Bible data achieves a substantial increase over the accuracy of a dictionary-only trained translator, and similarly happens when using two-step methods of fine-tuning. However, we also observed some, measurable, contaminated text from the Bible into the outputs of the best translator, creating concerns about its release to an Indigenous community. We end by discussing mechanisms to mitigate the negative impacts of this contamination.

This work explores the intrinsic limitations of the popular one-hot encoding method in classification of intents when detection of out-of-scope (OOS) inputs is required. Although recent work has shown that there can be significant improvements in OOS detection when the intent classes are represented as dense-vectors based on domain-specific knowledge, we argue in this paper that such gains are more likely due to advantages of the much richer topologies that can be created with dense vectors compared to the equidistant class representation assumed by one-hot encodings. We start by demonstrating how dense-vector encodings are able to create OOS spaces with much richer topologies. Then, we show empirically, using four standard intent classification datasets, that knowledge-free, randomly generated dense-vector encodings of intent classes can yield over 20% gains over one-hot encodings, producing better systems for open-world classification tasks, mostly from improvements in OOS detection.