Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The most hilarious thing about this is that Google has used "AI" to create the most un-inclusive tool of all, Google Translate.

I present you the beautiful translation of hungarian gender-less sentences to english:

ő csinos. ő okos. ő csúnya. ő jó.

becomes

she is pretty. he is clever. she is ugly. he is good.

I can't wait for inclusive warnings to come to non-english languages.



While it's hilarious, it's understandable IMHO - just basic statistics, she is more often used in front of pretty/ugly. It's showing a glitch in our society more than a glitch in Google AI.

This on the other hand looks like a hand-crafted blacklist of words that they want to remove from the language, I have no idea how would I train an AI which would classify "motherboard" as inappropriate.


> While it's hilarious, it's understandable IMHO - just basic statistics, she is more often used in front of pretty/ugly. It's showing a glitch in our society more than a glitch in Google AI.

Well, this doesn't necessarily show you facts about society because there's no particular reason to think the training set distribution is the same as anything in the real world.


You don't think it's reasonable to think the training set comes from the real world? Why not? Seems like a reasonable assumption to me.

The alternative is they are carefully curating a training set (google historically unwilling to do anything manually) or writing one themselves???


They said the training set distribution.

It's frankly impossible that they have a training set without any biases, though I'm sure they worked to eliminate the ones they could think of (which itself would have bias).


You can make the distribution different by eg duplicating some of the data a lot, which you might want to do to improve the actual purpose of the model (translation). Any other purposes (having opinions on gender) is just a coincidence and not being optimized for/regression tested.


> I have no idea how would I train an AI which would classify "motherboard" as inappropriate.

I believe you just need to give it emotions.


That's nothing

Ő jóképű

Becomes, wait for it

She is handsome

Just like in English, handsome is used to compliment a man, and pretty to compliment a woman.


if only English had a gender neutral pronoun that could be used by default when the gender of the subject isn't provided :thinking:


I still unironically want a singular variant to rise in popularity, don't particularly care what. There are so many times when "they" comes up and I have no clue if we're talking about a person or a group. There's already enough pronoun ambiguity when talking about single subjects without intoducing a dimension of plurality ambiguity.


If you think that's bad, you should see how ambiguous the plurality of the gender-neutral 2nd person pronoun is!


In some geographic areas, we have "you", "y'all" and "all y'all" which correspond to "tú", "vosotros", and "todos vosotros".


New Zealand (maybe Australia too?) has a solution for that in informal speech: yous!


We have one, but most people probably shouldn't use it.


Many folks I know use "they". Works well enough, I think.


Shall we write "they is pretty" or shall we wonder how many people we are talking about?


We shall wonder about the how many people. This is called the "singular they" and it has been a feature of English for hundreds of years.

https://en.wikipedia.org/wiki/Singular_they


Yes, I'm sure almost everyone reading this was already aware. This doesn't mean that "they are pretty. they are clever. they are ugly. they are good." is a good localization of "ő csinos. ő okos. ő csúnya. ő jó."


It's computer-brain solution to a problem that is crying out for a human. Humans "don't scale". Therefore paying 5 bi-lingual experts to teach the machine is "infeasible", but spending millions upon millions of dollars on hardware and software development, then exposing it the unwashed mess of the internet isn't....




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: