Your morality filters have arrived, dear peasants
Dedoimedo
A Place to Learn a Lot About a Lot
≡ Menu
Computers
Games
World
Art
Books
Cars
Physics
About
Back to Top
Your morality filters have arrived, dear peasants
Updated: August 5, 2026
A friend of mine sent me a link to a 9gag meme. The link had a preview thumbnail. I have my WhatsApp<br>client set not to show previews, and yet I could see it. Hm. Now, the thumbnail had the word<br>SENSITIVE written over it. I've never seen this before. And so my friend and I started a debugging<br>session, to figure out whence this nonsense comes. Very soon, we realized, with a solid 95% certainty,<br>that this was WhatsApp imposing its California-flavored morality onto us, two dinosaur Europeans old<br>enough to start a geriatric clinic. Yes, WhatsApp seems to be analyzing and labeling stuff based on<br>some outdated 19th century morality shit compass. We weren't quite sure, but the limited data and<br>testing pointed thus.
Morality wise, this ain't nothing new. After all, the whole world is galloping toward digital<br>control crap, to "protect" our feeble minds from dangerous content. Soon, everywhere, we might need to<br>"show our papers" like obedient monkeys every once in a while, to be allowed to use the Internet. Or<br>something along those lines, just to keep the peasants humble. The data fetishist worldwide will have<br>fresh troves and troves of identity information to play with. What I experienced the other day seems<br>like a precursor. Let's rant.
Figuring out the culprit
Before crying wolf, both of us decided to be quite diligent, and investigate what gives:
Both of us are running Android phones. When this happened, I was using the Web version of the<br>app through the browser.
My friend had link previews enabled, I did not. Once my friend turned off link previews, there<br>were no more thumbnails generated, and consequently, no labeling. But that does not tell me<br>anything about the actual link scanning and/or categorization done by whatever tool on the sender's<br>side. Also, in WhatsApp, the phone app and the Web version both have their own toggles for link<br>previews, and you need to set them separately.
We witnessed the labeling only for links shared by my friend while he had link previews turned<br>on. We did not witness it in the opposite direction, i.e., me as the sender and my friend as the<br>recipient. A possible reason, other than WhatsApp, is that I do not allow automatic system updates of<br>any kind, including Google services and related components, so my device hasn't been "blessed" with<br>the morality purification algorithms. Yet.
The labeling seems to be done by AI, and it seems to be fickle. Stuff we "expected" to be<br>labeled as sensitive, so to speak, wasn't always, and vice versa, totally silly things were blurred<br>and marked. We tested about ten random memes (all from 9gag, mind), and the only thing in common<br>was that, most likely, WhatsApp was doing something, based on some criteria (more on this<br>soon).
The label does not come from 9gag, though because some of the memes - most in fact - aren't in<br>the NSFW category.
WhatsApp blurs the preview when the label shows up (or is added, to be more precise), it does<br>not do so when there's no label. That would strongly imply it's WhatsApp doing things - or WhatsApp<br>is using something on the sender's phone (in this case, my friend's) before the content is<br>shared.
Furthermore, as of writing, the issue does not seem to happen with other chat clients.
And that's about it. That's the data that we have. A few theories come to mind. The mild version is<br>that WhatsApp (or something else Android-related) is scanning public links as part of its thumbnail<br>generation process, perhaps not unlike image proxying in mail clients. Perhaps. The semi-annoying<br>version is that WhatsApp (or whatever) is scanning, categorizing and sanitizing public links, for<br>whatever conceivable reason. The ultra-crappy version is that this is pure nannying and morality<br>reaching, and that I've just had a chance to see it in action. Now, whatever the reason, whatever the<br>culprit, the meme preview thumbnails were blurred and labeled. And there are two important aspects to<br>this phenomenon ...
Slow and steady idiot conditioning
Much like<br>QR code scanning, much like the upcoming<br>Android dev verification, this seems to<br>be another cog in the big idiot conditioning experiment that is the modern world post 'rona. Get the<br>plebs to behave nice like, cor. After all, scanning public links means the infrastructure for<br>... scanning is in place. If an app can decide to tell you whether a meme is "a-okay" before you<br>upload the link and send it to your friends in a private chat, then the said app can also do the same<br>thing with any which content. That means, it's only a matter of a switch flip before you're asked to<br>"verify" yourself before being allowed to see certain things.
One morality for all
Now, I hate politics, so I won't go too much into the whole collective social...