Publishing Partner: Cambridge University Press CUP Extra Publisher Login
amazon logo
More Info

Media: Popular Fully-Automated Translation Systems Have Gender Biases

Submitter: Brent Woo

Submitter Email:
Linguistic Field(s): Text/Corpus Linguistics

Media Body: This article discusses how the popular automatic machine translation services out there tend to have an extreme gender bias, due to the nature of using statistical analyses of existing corpora, which already exhibit such biases.

Read more:

(Editor's note: This editor asks whether or not this is simply a consequence of how certain natural languages, for example, those with grammatical gender, are used by humans anyways. In other words: is it really the machine's "problem" if people normally default to "engineer = [male] ingeniero"?)
Issue Number: 24.2258
Date Posted: May 31, 2013