Speciesist Language and Nonhuman Animal Bias in English Masked Language Models

Abstract

Warning: This paper contains examples of offensive language, including insulting or objectifying expressions.Various existing studies have analyzed what social biases are inherited by NLP models. These biases may directly or indirectly harm people, therefore previous studies have focused only on human attributes. If the social biases in NLP models can be indirectly harmful to humans involved, then the models can also indirectly harm nonhuman animals. However, until recently no research on social biases in NLP regarding nonhumans existed. In this paper 1 , we analyze biases to nonhuman animals, i.e. speciesist bias, inherent in English Masked Language Models. We analyze this bias using template-based and corpus-extracted sentences which contain speciesist (or non-speciesist) language, to show that these models tend to associate harmful words with nonhuman animals. Our code for reproducing the experiments will be made available on GitHub 2 .1 Anonymous previous version of this paper is accessible at https://openreview.net/forum?id=dfqMpjZOgv4, we have also published a paper on this topic in Japanese (Takeshita et al.

The ethics of sustainable AI: Why animals (should) matter for a sustainable use of AI

“…This can be argued for not just theoretically. Research on AI systems that are applied specifically to animals (Bossert & Hagendorff, 2021), as well as empirical investigations of speciesist bias in foundational language and vision models (Hagendorff et al, 2022;Takeshita et al, 2022) reveal the many disadvantages and harms that AI systems can afflict on animals. However, although a handful of papers demanding otherwise (Bendel, 2018;Owe & Baum, 2021;Ziesche, 2021), AI ethics has a purely anthropocentric tailoring where nonhumans are not part of normative considerations, despite the overlaps between concepts like fairness, well-being, sustainability, and so forth, with issues in animal ethics.…”