I got a notification of a page flag for a racist term.

The page in question was about home DIY - heating. The term in question was "wetback", which I had no idea was racist in the USA until someone at Google told me.

In Australia and New Zealand, a wetback is a special water heating device installed behind a fireplace. It's used in product names. It's even mentioned by New Zealand and Australian governments as part of energy saving initiatives e.g. [energywise.govt.nz...] and [level.org.nz...]

But explain this to a bot which takes no context or regionalisms....

Sometimes a scraper posts up scraped content as is - which could contain your code. I do not think that is still possible since they will only serve ads where your ads.txt is found - but I am not the resident AdSense genius - and scrapers might do just about anything.

Two ways to make sure that isn't the case is to search for identifying portions of your code within quotes to see whether it turns up on pr0n sites. Or you can log in at AdSense and make sure that only your domains are permitted to show your code.

I've exactly the same problem... I use adsense on a soft dating site and I keep get this messages everyday.
I ask for reviews, get accepted and the day after (ou 2 or 3 days), the page is back in the policy center, I review again, etc...

Since about one year I do that every morning...

@freitasm haha amazing. A company ran by bots, no employees to talk to. No surprise it gets everything wrong.

@not2easy adnsense policy center tells me the url and it's from my domain not some scrapper.

@Chewee Ok, i feel you. We got same issue. I'm gonna do a test, i'm not gonna send anything for review any more. I'm going to monitor it each day at same hour and take note of the number of urls, after a week i'll see if it grows or if google automatically clears the false positives (all of them). I'll let you know!

My guess would be a thumbnail in a side/end of page menu that randomly shows related content etc. That would explain random/multiple pages.

As for what's tripping the bot, mine was an illustration of a scientist with a goggle. With the right crop, the goggle was looking like boobs, and eyes like t..s

@Levo i think so too, i think it misinterprets images as sexual if they contain any number of skin tone pixels, even if they have nothing to do with naked humans.
Maybe the shapes too? i'm sure it's images, and i'm sure this bot is terrible at what it does, and google doesn't seems to care.
I also know this bot doesn't learn anything over time, cos it keeps flagging archives where i can tell which image got him confused, even after manual review clearance.

I'm waiting for them to lower the threshold on what this bot considers #*$!, because it's ridiculous now.

G's attitude is probably "better wrong than doing nothing at all..."

Which of course is still wrong ... just automated. :)

... less work to fix complaints than to police the web and pay someone to do that...

Just got an "Adult: Sexual Content" policy violation on a page covering naval aviation.

Maybe "naval" is too risque for the bot?

er ... maybe ... or is merely yet another indication that AI is not fully baked.

I also get these quite frequently. Usually it's an index page of products that's tagged, and since these are sorted by popularity the order of products changes all the time, so I never know which image (or word) could have triggered the violation. Oddly enough, those same words and images will also appear on the product page, but those are never flagged.

It would be helpful if they could screenshot the violation, or at least give some context beyond the URL.

It would be helpful if they could screenshot the violation, or at least give some context beyond the URL.

Won't happen. That would reveal the algo!

Maybe "naval" is too risque for the bot?

Clearly whoever programmed the bot knows the same rude version of "She'll be Coming Roud the Mountain" as me!

Luckily its not a problem that I ever hit on my own sites but remember emailing a friend at her office with a reference to "crackers" at our team Christmas lunch and having it blocked by her employer's profanity filter.

Test done, after 10 days of not sending anything for review > [i.imgur.com...]

My conclusion is, they never really go away entirely, some disappear on their own some new ones appear.
The only "solution" seem to be to keep sending them for manual review daily.
And if you happen to guess what the bot is picking up, change it (good luck).

Another day, another report of "Sexual Content" on a page that is far from it.

Wish they would apply this same sensitive "A.I" to their own ads; I've been seeing some very boundary-pushing suggestive ads coming through the network these days.

My main site is 17 1/2 years old. I had my very first violation warning in late 2017, for an archive post where someone had linked to a page that a screenshot of a G-rated Disney movie that showed the top of a man's butt crack. That page (on the third-party site) still has Adsense banners to this day, but me LINKING to it was a no-no.

After that first violation warning, I get at least 1 a week. Usually more. It's like, once they've looked at you, you're targeted forever.

It's hard to focus on building a business that relies on a third party for money, when that third party has the ability to cut you off, without warning, to no fault of your own.

Their AI is going nuts against some stop words used in the copy. Their strategy seems to be, restrict ad serving first, then have publishers request review to train the AI for further improvement. I get hit with sexual content consistently because I use the term "boobs" for a site of mine.

What boobies [en.wikipedia.org] could be... (Don't worry, SFW).

I just got another policy violation on the same article I got cleared for after requesting a review. It says dangerous and derogatory content, I have no vulgarity or offensive or nude content on my website

Ive had this blog for 12 years with no issue before, never received a policy violation until the last two or three weeks.

Ive received several violations.

Do you make enough to have email support - and have you contacted them for this or just clicked the button asking for a review?

Brace yourself, @Broaster... I was the same way, but once I got the first one I started getting them ALL the time :'-(

Two things...

1. Has your site traffic diminished?

2. These seem to be adjustments to your potential income (adsence, etc)

If your traffic has not diminished your site is still "in the serps", however your INCOME from g properties is likely imperiled.

Pick and choose what you want to do... censor yourself---or find another may to monetize.

Just know that in these days and times that diversity and "other voices" agenda might be part of the issue. We have a growing segment of the population world wide easily triggered by language that, only 10 years ago was (and remains) perfectly normal.

ON THE OTHER HAND, your competitors ... and there are many ... will do anything to knock you out so their offerings can increase.

We live in interesting times.