Which Words Would Google Consider Racist?

[csdude55]

Can anyone suggest where I might find a list of words and phrases that Google might consider to be racist? I've blocked everything I can think of from my message boards, but still get the occasional violation warning for something I'd never considered so I'd like to be proactive.

[Samizdata]

When problematic words/phrases are blocked, people just invent new ones:

Now a new type of hateful internet code appears to be emerging: The systematic use of innocuous words to stand in for offensive racial slurs. Search Twitter for "googles", "skypes" or "yahoos", and you will encounter some shocking results

Policing language is eternal mole-whacking.

...

[martinibuster]

I feel for you. That's like being caught between a hard place and a hard place, with rocks on either end.

Maybe a word filter isn't enough? Stepped up moderation? Like I said, it's a tough spot because increased moderation can kill a community used to a particular culture.

[csdude55]

Very true! Moderating is already a thankless 24/7 job that leaves me NO time to do any programming, but I don't have much of an alternative.

That's why I was hoping for an official list from Google, though. It's really not fair for me to get violation warnings for things I didn't know were a problem, and they seem to be going after user-generated-content more and more so I worry that one day I'll just get deactivated or something! A list that I could import regularly and just stay ahead of the game would be fair and much more practical for everyone.

Speaking of the impossibility of it all, though... "Here are all the innocuous things that suddenly became racist in 2017"

[nationalpost.com...]

[Sissi]

Unfortunately racism is not about words it is becoming more subtile
Reading between lines ist more hurting than some insults.
Racism is also linked to social discrimination .

[ember]

Almost anything one says today is considered racist by someone, somewhere. It is a losing battle, trying to police racist words.

[csdude55]

This is true, but since Google polices us, there should be some sort of current list so that we can prevent being penalized for words, phrases, and context that we didn't know that Google considered to be racist.

I think I can give this example. Back in 2007, Rush Limbaugh referred to Obama as "Magic N***o" (not the N-word... I don't want to post the word and get the site penalized, so Google it if you don't remember it). The reference was in news headlines, and someone copied a news article from HuffPost to my site in 2007.

I was recently given a policy violation warning for that article, even though it was in the news! And HuffPost still has the article on their site. I wouldn't have considered the word, phrase, or the context of the article to be racist, but Google did. So I'm hoping to find a list of things that I can or can't allow before getting further warnings.

[ken_b]

I doubt there is a public official, or even unofficial list of such words because providing one would only lead to never ending complaints that "that word isn't on the list".

[lucy24]

I was hoping for an official list from Google

Ha. Back in the '70's, give or take, there came a point where Czech writers were outright asking for censorship--that is, for an explicit directive spelling out what, exactly, they weren't allowed to say. Then they could make an informed choice about whether to say it or not say it, rather than learning only after the fact that they had offended. Evidently someone at G has decided it is to their advantage not to tell you ahead of time what's permitted.

Meanwhile, know your audience. (I cannot be the only American reading this site who finds endless merriment in the fact that the word "#*$!" is censored. But obviously it has meaning to somebody.) It's not enough pleasing the advertisers if the humans who were supposed to see those ads still avoid the site in disgust.

[piatkow]

It is impossible to keep up with everything that somebody will find offensive unless you are 15 years old with about 50 years experience.

[pubpolicycomms]

Hi everyone, and happy New Year. The Google UGC policy has been in place for many years, and we have articles on our Inside AdSense blog about this topic since 2009. A simple site search will back this up: [google.com...] I wrote about this myself in 2017: [adsense.googleblog.com...]

Users, and advertisers do not necessarily disassociate UGC from the content that you create for your site, thus the blog posts providing guidance on this topic since 2009, up until my most recent videos.

SEJ has posted two articles on this topic since late November, implying that this may be a new policy, which it is not. I am creating videos to help publishers better understand the reasons behind our policies, and this was simply a policy that was on my radar. Expect many more videos, on other policies in the coming months/year. We are not announcing new policies via video. These videos are meant to accompany our policy center, policy forum, our blog, and any other materials we create to help publishers better understand policies.

As for the specific ask for a list of terms, unfortunately we cannot make this public, but we are publicly stating in as many ways as possible that we are trying to support our ecosystem through these policies. And, I am trying to support publishers through materials that explain our policies. I hope this is helpful background.

[csdude55]

Very helpful, thank you. But now I have to ask, if I get policy violations for UGC even though I'm doing my best to comply, am I going to be at risk of having my account deactivated?

Without knowing what words and phrases are problems in advance, I have no way of pre-filtering them. I pre-filter the words and phrases that I personally know, but there are a lot more that I've never heard.

[pubpolicycomms]

We moved to page level monetization enforcement about 2 years ago now, so your risk is "generally" at a page level, not site/account level (unless egregious). Our recommendation, in my blog post from 2017 was to either enable real time monitoring, use a 3rd party monitoring solution, or place the comments after a call to action, such as "user comments", which open a new window without ads present. Many newspaper websites have moved to this model, as UGC can be difficult to manage, and users clearly have opted into this new "content" that they do not necessarily associate with your publisher site, and advertiser brands are not shown alongside questionable text.

[tangor]

All the above from @pubpolicycomms is regards UGC. With Alex Jones (a certified jerk from any point of view) and others recently banned there is some concern about content, UGC or otherwise. Today's environment on speech suggests a "list" or a more expressive policy more finely tuned rather than broad and vague specifications might be in order.

Meanwhile, commonsense and your gut will tell you what is reasonable or acceptable.

[edited by: martinibuster at 6:14 am (utc) on Jan 8, 2019]
[edit reason] off topic [/edit]

[martinibuster]

To expand on my post above, what I was getting at but left unstated is that a simple word filter might not be enough.

If someone denies the holocaust on your site, is it enough to filter out the word Holocaust?

So what I meant to communicate in my original post is that a word filter may not be enough. The solution may require an updated and more prominent posting guideline wrt to violence, racism and hate speech, followed with a respectful and polite approach to moderating the offenders. Given time the new policy may become a part of your forum's posting culture.

And that, the community culture, may be what needs a gentle nudge in the right direction.