Carsten Schwemmer, Carly Knight, Emily Bello-Pardo, Stan Oklobdzija, Martijn Schoonvelde, Jeffrey Lockhart (2020). Diagnosing Gender Bias in Image Recognition Systems. In Socius. URL: https://journals.sagepub.com/doi/full/10.1177/2378023120967171
In a new publication, the authors evaluate potential gender biases of commercial image recognition platforms using photographs of U.S. Members of Congress and a large number of Twitter images posted by these politicians. Their crowdsourced validation shows that commercial image recognition systems can produce labels that are correct and biased at the same time as they selectively report a subset of many possible true labels. They find that images of women received three times more annotations related to physical appearance. Moreover, women in images are recognized at substantially lower rates in comparison to men. They discuss how encoded biases like these affect the visibility of women, reinforce harmful gender stereotypes, and limit the validity of the insights we can gather from such data.