From The Guardian:
Researchers using AI technologies have discovered that male characters are four times more prevalent in literature than female characters.
Mayank Kejriwal at the University of Southern California’s Viterbi School of Engineering was inspired by work on gender biases and his own work on natural language processing to carry out the experiment.
Kejriwal and fellow researcher Akarsh Nagaraj used data from 3,000 books that are part of the Gutenberg Project, across genres including adventure, science fiction, mystery and romance.
The study used Named Entity Recognition (NER) to identify gender-specific characters by looking at things including female and male pronouns. The researchers also examined how many female characters were main characters.
“Gender bias is very real, and when we see females four times less in literature, it has a subliminal impact on people consuming the culture,” said Kejriwal. “We quantitatively revealed an indirect way in which bias persists in culture.”
But the researchers did face difficulties with those who didn’t fit into a gender binary. The AI was unable to figure out if “they” referred to a plural or a “non-dichotomous individual”.
Kejriwal said: “When we published the dataset paper, reviewers had this criticism that we were ignoring non-dichotomous genders. But we agreed with them, in a way. We think it’s completely suppressed, and we won’t be able to find many [transgender individuals or non-dichotomous individuals].”
As well as the statistics on male and female characters, the researchers also looked at the language associated with gender-specific characters. Nagaraj said: “Even with misattributions, the words associated with women were adjectives like ‘weak’, ‘amiable’, ‘pretty’ and sometimes ‘stupid’. For male characters, the words describing them included ‘leadership’, ‘power’, ‘strength’ and ‘politics’.”
Link to the rest at The Guardian