Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

vice.com

Post reply on HN
Print view View original

581–590 of 1001 posts

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

#581

Post by vorpalhex »

babypuncher wrote: post #579
vorpalhex wrote: post #372

Earlier quoted context omitted.

Being told I use too many adjectives, my word choice is too heavy for my audience or that I used too many clauses is useful. Being passive-aggressively told "blacklist" is a no-no word is not. If you were really clever, you'd let the user set a "sensitivity" level for the writing from "normal human speak" to "corporate PR approved".

In this particular case, the tool being discussed is meant almost exclusively for people writing "corporate PR approved" content, so it makes sense that the sensitivity setting is cranked.

I do a lot of things in Google docs beyond write PR releases...

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

#582

Post by CyberRabbi »

> Social editor Emily Lipstein typed “Motherboard” (as in, the name of this website) into a document and Google popped up to tell her she was being insensitive: “Inclusive warning. Some of these words may not be inclusive to all readers. Consider using different words.”

If blacklist is considered non-inclusive language, why isn’t motherboard? Motherboard is a needlessly gendered term that perpetuates stereotypes.

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

#584

Post by morgante »

Icathian wrote: post #292
meetups323 wrote: post #11

I wonder if this is an example of not knowing your audience (or rather assuming one's self represents a wider audience than it does). The people vouching for this at Google are likely product marketing managers, public relations folks, social media managers, etc. All they do is write corporate garbage all day, and much like we have "nit"s in PR's for formatting, variable names, etc; they likely have similar reviews t…

This explanation rings a lot more true to me than the rest of this thread. Everyone here seems to be looking into shadows for the woke gestapo, when I'd be willing to bet that this explanation is a lot closer to how this tool actually came to exist. I appreciate you adding to the conversation.

It's a little of both. There's enough woke influence over corporate communication to drive all employees to want to avoid non-inclusive words in docs. At that point, this becomes useful even for employees who don't agree with it—I'd rather just have the word flagged now and fix it instead of going back and forth later.

However, the net result is that words are driven out of the language even if everyone involved in the document wouldn't care.

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

#585

Post by matheusmoreira »

nicbou wrote: post #407
vorpalhex wrote: post #372

Earlier quoted context omitted.

Being told I use too many adjectives, my word choice is too heavy for my audience or that I used too many clauses is useful. Being passive-aggressively told "blacklist" is a no-no word is not. If you were really clever, you'd let the user set a "sensitivity" level for the writing from "normal human speak" to "corporate PR approved".

It depends on who you write for. As a non-native, non-American speaker, I'd be happy to know if I'm accidentally offending the people I'm writing to.

I'm not a native english speaker either. Chances are you aren't offending anyone. You probably know exactly which words of the english language are unambiguously offensive. Words like motherboard just aren't offensive.

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

#586

Post by julienb_sea »

I mean, this is the way of things. Corporate slacks have bots to correct non-inclusive language, and enforce syntax changes like whitelist --> allowlist, master --> main, etc. IMO it's silly, but nevertheless increasingly ingrained in corporate America. For the many of us that use G Suite for work, this feels like a natural extension of other areas to remind us to use current language. It's potentially a helpful reminder that avoids the awkwardness of someone actually making a gdocs comment about it.

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

#587

Post by titzer »

mgraczyk wrote: post #9

In the article they use the example of "Motherboard" and "Landlord" and seem to suggest that these are areas where the Google AI is making mistakes or being overly strict. As a Google employee expressing my own opinion and observations of company culture, I can say that these are 100% not mistakes. Many Google employees are just so out of touch with the real world that they believe it is the duty of Google Docs to ch…

Woke Clippy (to borrow from up-thread) is going to have a field day when it learns German. There. Are. So. Many. Violations.

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

#588

Post by pbhjpbhj »

matheusmoreira wrote: post #580
nostromo wrote: post #31

Who was asking for Woke Clippy? Is Google just bored and completely out of ideas? They seem increasingly disconnected with what their users actually want.

Yeah, who asked for this? One would think they'd be trying to achieve feature parity with Microsoft Office instead.

MS Word introduced a language censor like this recently (I've not seen it, only seen it announced at a workplace), so maybe they are following where MS leads?

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

#589

Post by nullc »

shadowgovt wrote: post #238
dekhn wrote: post #104

Earlier quoted context omitted.

I'm not worried about being Damore'd, as I have a lot more experience fighting progressives than he does.

The easiest way to not be Damore'd is to apologize and repent. Damore doubled down and at that point (because the gap between "manager" and "employee" at Google is so narrow) became a walking Title VII violation. Once his coworkers came out in public saying they wouldn't be able to work with him, Google was backed (legally and PR-wise) completely into a corner. It turns out American companies are not the Athenian Lyc…

> The easiest way to not be Damore'd is to apologize and repent.

What? Absolutely not. That is terrible advice when it comes to something that couldn't have been a literal accident. If he'd used the word "mother", then sure-- that could be apologized for. But a protracted essay on population level statistical differences between genders and its impact on the employment pool? Not a chance.

There is so much noise and outright disinformation about any issue that often the only reliable source for wrong doing is when the target of an accusation admits it themselves.

And even when that fails to protect you, at least you can be a hero to someone. Do you think a damore that apologized and said he was mistaken would be more employable? That people would eventually see it as a youthful transgression? I doubt it greatly-- it's not like the screens that show up when you google his name will yellow with age. Instead he'd just be the enemy to both factions of the war he wandered into, rather than enemy of one and hero to the other.

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

#590

Post by marton78 »

r_hoods_ghost wrote: post #77
twobitshifter wrote: post #48

Earlier quoted context omitted.

On landlord I can’t think of a proper synonym. Property owner and proprietor are broader categories of what a landlord is. They don’t mean the person you’re paying rent to for your housing.

rentier - one who lives off the income of property or securities.

Sounds like reindeer.
Post reply on HN