The paper "Language (Technology) is Power: A Critical Survey of ‘Bias’ in NLP" by Su Lin Blodgett critically examines 146 papers analyzing "bias" in natural language processing (NLP) systems. The authors find that these papers often lack clear and consistent motivations, normative reasoning, and engagement with relevant literature outside of NLP. They propose three recommendations to guide future research: (1) grounding work in the literature on the relationships between language and social hierarchies, (2) providing explicit statements of why certain system behaviors are harmful, and (3) examining language use in practice by engaging with the lived experiences of affected communities. The paper also includes a case study on African-American English (AAE) to illustrate these recommendations, highlighting how existing NLP systems may reproduce racial and linguistic biases.The paper "Language (Technology) is Power: A Critical Survey of ‘Bias’ in NLP" by Su Lin Blodgett critically examines 146 papers analyzing "bias" in natural language processing (NLP) systems. The authors find that these papers often lack clear and consistent motivations, normative reasoning, and engagement with relevant literature outside of NLP. They propose three recommendations to guide future research: (1) grounding work in the literature on the relationships between language and social hierarchies, (2) providing explicit statements of why certain system behaviors are harmful, and (3) examining language use in practice by engaging with the lived experiences of affected communities. The paper also includes a case study on African-American English (AAE) to illustrate these recommendations, highlighting how existing NLP systems may reproduce racial and linguistic biases.