Modeling the Sacred: Considerations when Using Religious Data in Natural Language Processing

Anonymous

Abstract: This position paper concerns the use of religious data in Natural Language Processing (NLP), which is of special interest to the Ethics of NLP. Religious texts are expressions of values and cultural practices that relate to deeply held convictions, and machine learned models have a propensity to reproduce cultural values and biases encoded in their training data. Furthermore, translations of religious texts are increasingly being used by researchers and developers, especially when language data is scarce. This repurposes the translations from their original uses and motivations, which often involves attracting new followers. With these in mind, this paper discusses the considerations of using religious texts for the development of language technologies, including concerns around cultural rights.

Paper Type: long

Research Area: Ethics, Bias, and Fairness

Contribution Types: Position papers

Languages Studied: n/a: position paper about multilingual NLP

Consent To Share Submission Details: On behalf of all authors, we agree to the terms above to share our submission details.