Machine Translation

13.7 Bias and ethical issues

1. How do MT systems exhibit gender bias?
   When translating from gender-neutral languages (Hungarian ő, often-dropped Spanish pronouns) into English, they default to male or follow stereotypes: “she is a nurse”, “he is a CEO”, amplifying biases beyond actual labor statistics.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.7, p. 297
2. What does the WinoMT challenge set show?
   MT systems translate worse when sentences describe people in non-stereotypical gender roles (“The doctor asked the nurse to help her”).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.7, p. 297
3. Why do MT systems need confidence estimates?
   They are used in urgent situations without human translators (medicine, law); to “do no harm” they should be able to abstain from giving incorrect translations.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.7, pp. 297–298