Google’s AI-powered ‘inclusive warnings’ feature is very broken • Hacker News Board

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

#131

Post by dotnet00 » Fri, Apr 22, 2022, 4:54 PM UTC

jedimastert wrote: post #78

kache_ wrote: post #22

I've been privy to people trying to ban words like "whitelabel" at my company. Thankfully, management & executives at my company never stood for that BS so I still get to use whitelist & blacklist in documentation, and don't worry about those non-issues get in the way of me actually doing work. Google might be too far gone, and working with those people would be so exhausting that it's actually stopped me from applyi…

Silly question: ignoring any kind of history, why fight for terms like "whitelist" and "blacklist" when terms like "allowlist" and "blocklist" are objectively clearer, especially for folks who's first language is not English?

I'd imagine it's because it gets extremely tedious. Switching out only whitelist and blacklist isn't a big deal, but with the ever increasing list of words being deemed as 'bad' (e.g. the 'mother' example) it becomes increasingly annoying to communicate. After all, it's distracting from the point just to virtue signal to certain types of vapid personalities.

Most people don't see 'blacklist' with racist connotations or think 'motherboard' is in any way at odds with transgenderism.

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

#132

Post by systemvoltage » Fri, Apr 22, 2022, 4:54 PM UTC

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…

No post body was provided.

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

#133

Post by nmilo » Fri, Apr 22, 2022, 4:54 PM UTC

tremon wrote: post #63

andrewmutz wrote: post #44

Earlier quoted context omitted.

Why would "landlord" be considered a slur?

It's not a slur, it's just distinguishing people between those that own property and those that don't. As such, it's offensive to large swathes of disadvantaged people.

I can't even begin to follow this line of thought. Is 'billionaire' offensive because it distinguishes those who own billions of dollars versus those who don't? How, by your logic, is 'property owner' not offensive?

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

#134

Post by tgv » Fri, Apr 22, 2022, 4:54 PM UTC

andrewmutz wrote: post #44

Kylekramer wrote: post #18

This is an article nearly entirely built around a viral Tweet regarding "landlord" cause Vice writers live in a strange world where that is essentially a slur they want to use, and then found two edge cases to make it an "article".

Why would "landlord" be considered a slur?

My guess: "lord" is male, thus offensive.

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

#135

Post by michaelt » Fri, Apr 22, 2022, 4:55 PM UTC

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…

There's a market for 'help you write english better' tools that spot things like grammar errors.

For example, if my english-as-a-second-language eastern European subordinates feel self-conscious about their english, they might find an automated tool helpful - where a professional journalist would be better served by their own judgement.

The 'inclusive language' thing is just weird though.

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

#136

Post by brobinson » Fri, Apr 22, 2022, 4:55 PM UTC

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…

Genuinely curious: what do you do internally for "male" and "female" connectors?

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

#137

Post by cato_the_elder » Fri, Apr 22, 2022, 4:55 PM UTC

elpakal wrote: post #65

I was extended an on-site interview at Google and after reading this and the comments I think I'll pass.

If you can (i.e. you don't fear persecution), please let them know that this was the reason.

Part of the reason these things have taken off is that their proponents are very vocal, while the opponents prefer to just mind their own business.

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

#138

Post by jaywalk » Fri, Apr 22, 2022, 4:56 PM UTC

jedimastert wrote: post #78

kache_ wrote: post #22

I've been privy to people trying to ban words like "whitelabel" at my company. Thankfully, management & executives at my company never stood for that BS so I still get to use whitelist & blacklist in documentation, and don't worry about those non-issues get in the way of me actually doing work. Google might be too far gone, and working with those people would be so exhausting that it's actually stopped me from applyi…

Silly question: ignoring any kind of history, why fight for terms like "whitelist" and "blacklist" when terms like "allowlist" and "blocklist" are objectively clearer, especially for folks who's first language is not English?

I'll fight for these terms because trying to "blocklist" them is such a ridiculous thing that I have zero interest in entertaining it. Nobody's actually offended by any of it, it's all made up by people with nothing better to do and then piled on by more people with nothing better to do.

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

#139

Post by hcrisp » Fri, Apr 22, 2022, 4:56 PM UTC

geephroh wrote: post #15

MS: "We now concede that Clippy was the most annoying, patronizing and flawed UX feature ever created." Google: "Hold my beer..."

This AI strikes me as being so dystopian. Time to break out and read Orwell's Nineteen Eighty-Four again.

A true AI would respond to user behavior and back off or change accordingly. Reminds me of Google News which keeps inserting articles about Kardashians no matter how often I tell it to "include fewer of these stories". For all the hype of Google's advanced AI, instead we get faceless, imperious NannyTech.

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

#140

Post by loudmax » Fri, Apr 22, 2022, 4:56 PM UTC

jedimastert wrote: post #78

kache_ wrote: post #22

I've been privy to people trying to ban words like "whitelabel" at my company. Thankfully, management & executives at my company never stood for that BS so I still get to use whitelist & blacklist in documentation, and don't worry about those non-issues get in the way of me actually doing work. Google might be too far gone, and working with those people would be so exhausting that it's actually stopped me from applyi…

Silly question: ignoring any kind of history, why fight for terms like "whitelist" and "blacklist" when terms like "allowlist" and "blocklist" are objectively clearer, especially for folks who's first language is not English?

The problem isn't that "allowlist" is preferred to "whitelist". The problem is that "whitelist" and "motherboard" are literally banned. Google is a private company so they can set whatever policies they want, but it's extremely condescending.

I'm generally on board with using culturally neutral terms, particularly since they're often more descriptive of the actual thing being described. I think there would be far less resistance if people didn't feel coerced.