Live data from Hacker News

Google’s AI-powered ‘inclusive warnings’ feature is very broken

vice.com

951–960 of 1001 posts

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#951

This is straight out of Brave New World where the word "mother" was viewed as obscene.

Check out this video uploaded just recently: [Health department refuses to define ‘woman’ in Senate Estimates]( https://www.youtube.com/watch?v=DX_1QNXgjDM ). Australia's own health department struggle and eventually refuse to answer the simple question. Look at how uncomfortable they are with such an innocuous request. On similar trends, we are seeing institutions such as the ACLU post tweets like this: https://twit…

> Look at how uncomfortable they are with such an innocuous request.

Ah, but it's not innocuous, and everyone knows it. I assume you know it -- when you watched that clip of the senator, did you really think he had forgotten what the definition of "woman" was?

The question he was really asking was along the lines of "is it the health department's official stance that trans woman are woman and should be treated as such under the law?"

Although actually he wasn't really asking that either, he was signaling to his supporters that he does not think that trans woman are women, and he was attempting to embarrass his political opposition so he and his allies could talk about the lunacy of the woke left or whatever and how they can't even define what a woman is.

I can forgive people for not wanting to play along.

Although, I can't forgive people for not having an answer ready; that question is being used as a "gotcha" more and more frequently.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#952
post #869

Earlier quoted context omitted.

> 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???

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.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#953

Virtue signaling is a hell of a drug. I tried to explain to someone online that the word black has many meanings, not every use of black is about people. Oh man, in the end I was "taught" that just by not accepting his ideas I was a racist person. My experience with this type of person is very bad. They are ignorant people, defending points without understanding the points. They are very aggressive and work in packs.…

When I first met the D&D alignment chart, it was enlightening. It can explain a situation like this so easily - namely, that the person you met is a lawful evil type. Like a concern troll on the internet, bad intentions are hidden behind a legal, orderly facade.

I don't think these types should discredit the movement. I do wish they wouldn't rile up folk so much though. Change is hard as it is, it's very unhelpful for the cause that they behave in such a bad way.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#954

The good news is that those who are against this stuff may be pleased to know you're not alone --- roughly half of the Americans on one side of the political spectrum would agree with you. The bad news is that Google and seemingly the rest of Big Tech are all sitting solidly on one side, and amongst all this "diversity and inclusion" talk, there is basically no thought given about diversity and inclusion of the polit…

Google is a private entity. In the opinion of the economic right, wouldn't companies be free to make their own rules about their own conduct?

Big tech is not how you're painting it. I think it sits solidly where it can keep its income growing. They'd like to operate in China, so they apply filters to search results, or use their servers instead of the default. Pride month sees avatars changing to a colorful, but only in the more permissive locations. Business is, in part, about getting along, and inclusivity is currently how you get along in the West - nothing more, nothing less.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#955
post #621
post #242

Earlier quoted context omitted.

"Blacklist" and "whitelist" suck anyway because they require the cultural knowledge that "black is bad and white is good" to understand their meaning. "Allowlist" and "denylist" don't require any cultural background to understand - their names are purely descriptive. They are just better terms.

Blacklist has nothing to do with skin-color...you fall in the same trap as those wannabe corrector's: >>His memory was stored with a black list of the enemies and rivals, who had traduced his merit, opposed his greatness, or insulted his misfortunes https://en.wikipedia.org/wiki/Blacklisting#Origins_of_the_te...

Excuse me? I didn't claimed it did. Nor do I think it does.

Like what the actual fuck? I sincerely hope this was meant to be a reply to someone else, otherwise you need to drastically improve your reading comprehension.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#956

The comments on this post sure are something. Do I really have this many peers that apparently live in constant fear of the woke Boogeyman?

I think what can be seen is people feeling like someone forcing a change on them. Language, culture, habits are really personal. I think the less someone thinks about theirs consciously, the more personal they are.

Also, as many point it out, this change is often spearheaded by entities of questionable reputation. Google, for one, is grilled a lot for their behavior lately, and now they roll this out.

Lastly, I think that this innate and natural feeling, that "they" will come and change our ways, is preyed upon by populist politicians. When people feel like this, it can be, and is being played up by creating the very Boogeypersons you jokingly made up: they point to immigrants, like LGBT people, successful Jewish businessmen, whatever sticks.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#957

Of master is not acceptable anymore, how are we going to call the MSc or MA degrees?

SAP, for one, makes an exception for the word "master" for cases like the Master's degree, or Scrum Master:

https://help.sap.com/doc/b0322267728e48a28b0c8ee7dd1ab4c7/1....

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#958
post #276

I'm going to delete my Google account today. I'd rather watch YouTube without ads, but I can take a little brainwashing for some praxis. Corporations only speak the language of money, outrage and morals are foreign to them.

Come on over to Onlyoffice, they have an online office offering and no inclusive suggestions.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#959

Earlier quoted context omitted.

>and neither did my ancestors Considering the number of ancestors in your past, this is highly unlikely for any human being alive today.

If every human being alive has had a slavemaster ancestor, all the more reason to not care about the use of the word in a completely unrelated context involving hard drives or branches of information.

If we know that a behavior is harmful, all the more reason to stop keeping it up. I'm sure we can manage technology without edgy analogies to people fucking or suffering.

Re: Google’s AI-powered ‘inclusive warnings’ feature is very broken

#960

Earlier quoted context omitted.

Your question takes for granted that people are fighting for whitelist in blacklist, but it's the opposite. Whitelist and blacklist are long-established and familiar terms.

No, they are not. Teach someone with no previous cultural knowledge what "allow" and "deny" means, then ask them what white list and black list means, then ask them what allow list and deny list means. Black and white list requires previous knowledge of what those terms mean, allow and deny do not. Someone can intuit from just basic language knowledge what the feature does.

My pet peeve is that the new, inclusive terms are often more cumbersome. For example, allowlist is longer by a syllable. I know that it's hard, but I wish that people came up with terms that are just as handy as the old ones. Changing blacklist to blocklist is quite cool for one.
Post reply on HN