-2

Note: this somewhat related to my previous question

This article deals with what seems to be gender bias in machine translation:

"Our results show that male defaults are not only prominent but exaggerated in fields suggested to be troubled with gender stereotypes, such as STEM (Science, Technology, Engineering and Mathematics) jobs," the paper says.

Further evidence of algorithmic bias – which might be described as failure to compensate for cultural favoritism – showed up in the associations of certain adjectives with certain gender pronouns. Sentences with the words "attractive," "ashamed," "happy," "kind," and "shy" tended to be translated with female pronouns. Sentences with "arrogant," "cruel," and "guilty" were translated as male.

What's more, the researchers speculate that the bias shown in English may influence other languages, because "Google Translate typically uses English as a lingua franca to translate between other languages."

Some of commentators argue that the algorithm itself is not biased, but the corpus used to learn from is:

It's a machine learning algorithm, it learns translations from a corpus of texts. Its gender bias represents the bias of the corpus.

I am wondering if this topic was or is mentioned in US politics.

Question: Is there a tendency, in the United States, to ask for altering machine translation algorithms to use gender-neutral language and gender pronouns?

Note: As a personal note, I find the intersection between ideology and algorithms particularly interesting.

3

Not yet...

Insisting that machine learning algorithm use gender neutral language very much a niche topic. Your question is the first I've heard of it.

... but what "machine learning algorithms" do and their relative bias is an emerging field of discussion

What is much more frequently being discussed is how machine learning algorithms, generally, might be biased against or for certain types of people.

The most prominent example of this sort of discussion comes up with regard to facial recognition and matching technology. Adequate data sets for training face matches based on machine learning are actually very difficult to come by, so many companies working on this use photos of their own employees. Software companies tend to be staffed with white and/or asian men, so the resulting algorithms tend to do better at identifying such people than they are at identifying women or people from other ethnic groups. This is starting to become controversial politically because police departments wish to use face matching, and there is already political controversy (in some cases going back before the founding) about how minorities are disproportionately affected by law enforcement activities and the reasons why. So, there is concern that machine learning might make these existing concerns worse in various ways and discussion ongoing about how to deal with that, possibly through legislation.

So yeah, I wouldn't say that the issue you're asking about is an actual political movement yet, but it will probably come up if the current thinking on gender-neutral language persists for a few more years at the same time as other discussions about regulating machine learning algorithms are taking place.

| improve this answer | |

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .