Language (Technology) is Power: A Critical Survey of "Bias" in NLP

TL;DR

This paper critically surveys 146 NLP bias studies, highlighting issues in motivation, methodology, and normative reasoning, and proposes recommendations for more socially aware and responsible bias analysis.

Contribution

It provides a critical analysis of existing NLP bias research and offers guidelines to improve normative reasoning and community engagement in bias studies.

Findings

Motivations for bias analysis are often vague and inconsistent.

Current quantitative techniques do not align well with stated motivations.

Recommendations emphasize normative reasoning and community-centered approaches.

Abstract

We survey 146 papers analyzing "bias" in NLP systems, finding that their motivations are often vague, inconsistent, and lacking in normative reasoning, despite the fact that analyzing "bias" is an inherently normative process. We further find that these papers' proposed quantitative techniques for measuring or mitigating "bias" are poorly matched to their motivations and do not engage with the relevant literature outside of NLP. Based on these findings, we describe the beginnings of a path forward by proposing three recommendations that should guide work analyzing "bias" in NLP systems. These recommendations rest on a greater recognition of the relationships between language and social hierarchies, encouraging researchers and practitioners to articulate their conceptualizations of "bias"---i.e., what kinds of system behaviors are harmful, in what ways, to whom, and why, as well as the normative reasoning underlying these statements---and to center work around the lived experiences of members of communities affected by NLP systems, while interrogating and reimagining the power relations between technologists and such communities.