<p>The history of the study of emotions has given way to several disciplines that have expanded our knowledge of the human mind, like psychology and philosophy. In our current times, Sentiment Analysis presents itself as a new path to study this field through automated and computerized means. However, automated systems can also carry many biases, such as sexism, thus possibly skewing their analysis towards a perpetuation of misogynistic views. By employing different Sentiment Analysis models on two sexism inventories, we observe that the emotional analysis these methods carry out varies in effectiveness and that this analysis can only reach a superficial understanding of sexism. Through these findings, we elaborate a critique of the technological concept of objectivity and neutrality in science. This study brings the feminist epistemological view to light to warn against an uncritical, indiscriminate use of technologies that can result in the further perpetuation and standardization of sexist biases in science.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Women are too easily offended, but robots aren't: a feminist critique of sentiment analysis

  • Darío Doña-Falcón,
  • Pilar Medina-Bravo

摘要

The history of the study of emotions has given way to several disciplines that have expanded our knowledge of the human mind, like psychology and philosophy. In our current times, Sentiment Analysis presents itself as a new path to study this field through automated and computerized means. However, automated systems can also carry many biases, such as sexism, thus possibly skewing their analysis towards a perpetuation of misogynistic views. By employing different Sentiment Analysis models on two sexism inventories, we observe that the emotional analysis these methods carry out varies in effectiveness and that this analysis can only reach a superficial understanding of sexism. Through these findings, we elaborate a critique of the technological concept of objectivity and neutrality in science. This study brings the feminist epistemological view to light to warn against an uncritical, indiscriminate use of technologies that can result in the further perpetuation and standardization of sexist biases in science.