When we think of “people” do we tend to think of men?

Anyone who tries to use gender-neutral language often uses gender-neutral terms such as “people”, “people” or “people”. But how neutral are these terms really? An analysis of over 630 billion English-speaking words from the Internet has shown that even seemingly neutral terms are most often used in a male-related context. This suggests that when we hear terms like “people” or “person,” we are thinking of men rather than women.

What are the hidden concepts behind the words we use? To find out, artificial intelligence analyzes were developed. Artificial Intelligence evaluates which words are used in a similar context and therefore likely have a similar meaning. For example, “Joe puts on a balak to boil water for tea” would mean that “balak” has the same meaning as “kettle”. Similarly, such analyzes have already shown that, for example, there is a high correspondence between the words “scientist” and “researcher” and a higher correspondence of these words with the word “clever” than with the word “instead of”.

How gender fair are the neutral terms?

“Many forms of bias have been explored in this way, such as the tendency to associate” science “with men rather than women, explains April Bailey of New York University. “On the other hand, there has been hardly any work on how we perceive the” person “.” She and her team have now addressed the question. To this end, they used a dataset containing more than 630 billion English-speaking words used on nearly three billion websites.

In three sub-studies, Bailey and her team tested the consistency of neutral terms such as “people” with gender-specific terms such as “male” and “female” and their consistency with adjectives and verbs related to masculine or feminine. “Our results show that even when we use gender neutral terms, we prefer men to women,” said co-author Adina Williams of Facebook Artificial Intelligence Research in New York.

“People” with male characteristics

The first sub-study found that words like “people”, “person”, “human” and “someone” are used in a similar context for clearly masculine gender assignments such as “male”, “male”, “male” and ” Hi”. On the other hand, there was much less agreement for female assignments such as “Woman”, “Woman”, “Woman”, and “She”. “This was also true when we excluded the word” human “from the analysis, which could also be generally used as” human, “the team of authors said.

In the second sub-study, they focused on adjectives related to specific gender stereotypes, such as “empathetic”, “family-oriented” and “friendly” to women, and “energetic”, “rational” and “controlling” for men. Using machine similarity analysis, they showed that these attributions were, in fact, often used in appropriate gender contexts. Now they have analyzed the similarity between adjectives related to men and women with gender neutral terms such as “People” – and indeed: here, too, it has been shown that “People” is used more often with adjectives related to men than those related to women.

men preferred

The third sub-study in which verbs were used instead of adjectives (including “admire”, “complain”, “kiss” for women and “honor”, “respect”, “kill” for men) came to the same conclusion . ‘These results show that the authors of the analyzed Internet texts write (and to some extent presumably also think) more similarly about men and men than about men and women,’ the research team concludes. “This suggests that the collective concept of people favors men over women.”

From the point of view of Bailey’s colleague Andrei Cimpian, this is socially questionable. “People’s ideas” form the basis of many social decisions and political measures, he says. “Since both males and females make up about half of our species, our shared notion of a” person “favoring males leads to unequal treatment of women in decisions based on that concept.” it is important to be aware of this unequal treatment and to take measures, for example when programming future AI, to avoid such distortion.

Source: April Bailey (New York University, USA) et al., Science Advances, doi: 10.1126 / sciadv.abm2463

Leave a Comment