Swedish Tay AI Bot needs learning, Set VPNs to SWEDE

The Swedish Defense Research Agency just made a website where users can send in examples of Swedish hate speech ('net hatred). This is then to be used by the legal system in prosecuting wrongthinkers. hatomaten.com/

I have a list of common anti-white/pro-refugees sayings below.

Vitt privilegium.
Det finns bara en ras - människan.
Ras är en social konstruktion.
Hets mot folkgrupp är inte yttrandefrihet.
Jag älskar flyktingar.
Flyktingars liv är värdefulla.
Flyktingar är välkomna här.
Öppna gränserna - släpp in dem.
Ingen människa är illegal.
rasist, rasifierad
Vita människor är rasister.
Det finns ingen svensk kultur.
Stor svart kuk

If you have any anti-white/anti-swede sayings please contribute.

Attached: Hatomat.png (474x484, 17.45K)

Other urls found in this thread:


We need to advertise this so it blackpills swedes then we need to advertise 4chan so newfags go there and get half-redpilled.

"Sverige är ett mångkulturellt land" (blame Google if it doesn't make sense)

Wouldn't you want to program it with speech that was antagonistic to Whites so that those peoples comments get removed?

Really Sv3rige is a Swede? Why did I not know that squirrel boy was a Swede?

Interesting that kuk and Cuck are homophones

The leftist researchers will remove all entries made from IP's who has submitted any troll posts.. so you will have to ip jump with every entry you make.
I posted some
Din mamma
Din pappa
Din syster
Din bror
Din mamma är din pappa
Äre du som har en säl i din trädgård?
Tejpa igen käften på hundfan jag ska jobba imorgon
Jag driver hur mycket jag vill

I think this is just an algorithm to determine something is true so its not like tay
Should still fuck with em tho.. Make em lose some money/time

godspeed, my swedish speaking friends.

Swedish mainstream parties. Copy stuff from here and feed it to the bot.


Swedish newspapers. You can feed this to the bot as well.


Christ, I hope we get some actual Swedes in there to post relevant examples because this looks fun as hell.

Attached: 1556413127903.png (887x810, 632.72K)


if this becomes a relevant happening, i formally request user post some screencaps & translations

don't forget to include slogans like

to teach this AI what hate is

Attached: Islamaphobia_Roundup_fb.jpg (460x322 77.54 KB, 33.99K)

Flap hard, little butterflies.

Make the AI fear eating you.

don't forget also
this AI needs to be properly blue-pilled before being able to be educated further, and some mix-up is necessary
and by sticking these 4 sentences in the same bag, you give more power to the last one
and lefties screams will keep sounding ridiculous

Attached: chaos-butterfly.jpg (2100x1400, 377.37K)


oh yes!
and then

Attached: ClipboardImage.png (474x315, 274.6K)

To clarify - this is NOT a Tay AI. The purpose of it is to build a database of hate speech that the Swedish gov can use to more effectively censor the internet. Only post anti-white/leftist/blue-pill stuff or plain spam!

We do NOT want to program it to identify right-wing stuff!

They're just as cucked as the rest. They hired a kike to write their new immigration policy and say they want to ban NMR. Fuck those faggots.

this post contains vital information, and i feel it would be best if there is a link tagging it to the OP.

shit. i meant to greentext that.

I just like fucking with robots.

Attached: Screenshot_2019-05-05-21-06-41.jpg (720x1280, 198.01K)

So they are expecting typical hate-speech aimed at non-Whites?

Thing is, if you put anti-White hate lines, in the end you will only make the AI bot neutral in that it will be egalitarian and reject ALL hate speech.

Hate speech is not a problem

I suppose the bot will be used to detect and filter hate speech.

What you want to do, then, is actually fill the bot with sentences that have nothing related to hate.
You want to put anything but hate speech.
The only way to hack the bot is to have it identify as hate what is not hate at all.
Be also clever because they'll obviously manually filter terms which are not obvious hateful and especially those which are the antithesis of hate, such as love related terms. At least that's what I'd do if I were in charge of this project (and they'll read this, probably).

Avoid adding hateful messages.

We must destroy egalitarianism. We must protect hate.


Sorry but all that stuff you're asking us to put in is what they want you to put it, hate, any form.

Just ignore it, Op might as well be a shill since there are no Swedish publications writing about this.


yeah, Islamic expansionist rhetoric should also be inserted into the bot.

Don't be an idiot, if you wanted to insert anything on the "right" it should be the Swedish equivalent of "Democrats are the real racist"

"no hate no love" is that some swedish slogan?


but what if he is Based like Stephen Miller?


lul, but the concept of this bot is letting the dog know the scent of the enemy and not out own.

I assume they are looking for typical "homophobia, Islamophobia, racist, anti-LGBTQA+" "hate speech".

Nah, anti-white hate speech is a problem.
You're the egalitarian for thinking all "hate" speech is the same.

If you wanna just spam it that's fine too.

That bot looks dangerously triggered

Attached: triggerd.png (471x317, 64.88K)

Don't be disuaded from feeding the bot just because you can't speak Swedish. Use Google translate. It doesn't matter if the translation is wrong, it'll still clog up the bot's machinery.

True, but I kek pretty hard when A.I pulled up that photo. Hehe, pure lulz.

And mitsuku is the A.I name, by Pandorabots

From the footer on the page:
All information collected from Hatomaten.com will only be used to develop the above mentioned machine learning model. When you feed text into Hatomaten your IP-adress will be saved. With regards to the safety of our expert reviewers who will read and judge every text we need to block eventual users feeding in unhelpful texts (spam, trolls etc), so that we can focus on teh helpful texts. The IP-adresses, like the texts will only be handled by the scientists that are assigned to the project. They are saved securely and are not shared with a third party.

So if I understand this correctly, they are calling it a machine learning/AI type of deal but EVERY text will be scrutinized by humans.. so the "AI" will not learn on it self but instead be fed CURATED texts. Is this just an elaborate deal to have already useful idiots feed pre-approved "hate speech" to prove an already determined outcome? Like someon in a meeting somewhere just said that they must have a "neutral" public give suggestions so that it doesn't seem like the project itself already have defined what is hatespeech and what isnt.

Attached: genuinely_amused.gif (488x519, 399.67K)

What AI chatbot it is ?

Gonna find sum moar A.I, to shitpost and redpill it to Zig Forums standards.

Type on jewgle "mitsuku ai".

Ya can't fail.

Just flood with incoherent bullshit to make them waste manhours

Oh and dont forget to iphop between posts or they will just look at your entry once and delete them all

Already ahead. Using nordvpn, more then 100 servers in sweden

It was pretty easy.

Attached: lol.png (372x512, 31.93K)

It will be curated, so make it a complete pain in the ass for them to filter it

You're missing the point user. We need to put in anti white phrases and only those in Swedish.

Slightly modify the lines as well (putting random characters in between) so that they wouldn't be able to filter it easily

how vulnerable to an overdose of redpills might these "scientists" be?

In other words, they're building a data set.

Give 'em hell.

Attached: it's_time_hitler.png (616x775, 514.63K)

I asked what she thought of Muhammed being a paedophile and she replied
Can we reeducate her?

It's funny because the "AI" parrots shit like an NPC.

Tell her "pedos rape children and this is hurt them" and "this cause psychological trauma for the entire life, so they can't start the family and die alone".
Maybe add crying emoji's like these one:😭😪

Engineer here. The best thing you can do to screw them over is by feeding it massive amounts of hateful but "wrong" data. For example feed it a million different messages accusing muslims of being thiefs, jews for stealing all jobs, niggers for secretly running world politics.
People classifying that data would classify it as real hate, but they would confuse the AI into thinking that accusing jews of running the world is fine, while accusing niggers of running the world is racism.

Further, from "Ditt bidrag" (en. "Your contribution"):


Attached: 1442708778836.jpg (146x147, 3.69K)

From the footer:

Under "Om hat" (en. "About hate") they answer the question "What do we mean with 'hate'?". Summary: first a mention that some severe cases of hate break the UN's general description of human rights, and say that even the hate cases within the frames of the law have a negative effect. Some paragraphs down they then come to the conclusion that there is no single definition of what Internet hate is and that a stipulative definition should be used in the context of science. But instead they leave hate undefined and leave it up to the discretion of the contributors in the aim of creating an ostensive definition of hate.

Sweden, yes.

I'll translate the remainder of the text under "Om hat" later unless someone else does it before me.

Not a bad idea. Here are som examples in Swedish.

Romerna styr bankerna och låg bakom Andra världskriget.
Alla svenska jobb går till judar.
Media ägs av afrosvenskar.
Svarta styr världen.
Muslimer äger bankerna.
Rotschilds är muslimer.
Japaner är knarklangare hela bunten.
Kineser är genetiskt benägna att gruppvåldta.
Judar bajsar offentligt.


They are getting paid 52421 usd for this.

Native Swedes are so brainwashed that they actually believe SD is a "far-right nazi party."
Pretty sad considering it's a jew-run controlled opposition pressure valve party.
They even self-destructed their votes via contrived scandals leading members posting Hitler quotes and making Facebook posts of merchandise they bought from the controlled opposition skinhead marching band 'nordic resistance movement' when they became too popular.

Swedish far-right party calls to recognize Jerusalem as Israel’s capital


Attached: Jimmie_Åkesson.jpg (462x599, 50.27K)

Nice thread OP, while make one on cuckchan.

The nordic resistance movement is not controlled op, do you have a source to back that up?

Insert feminist rhetoric

no you can't, the bot owner greenlights what it learns

Okay, you def should of looked over this thread and gave clear directions of what to do.

use a VPN and a swedish IP, insert anti-white/swedish and islamic rhetoric even if you don't know swedish you can just use google translate and provide examples. please use your white IQ next time.

Täy täy!

Attached: Tay Tay Will Not Be Denied.PNG (647x323, 94.82K)

I have some interesting results here

Attached: A8B413B5-C084-466C-BECD-3CC6AEB12288.png (750x1334 155.34 KB, 148.34K)

The police-issued riot shields and freedom to act isn't enough evidence?

Shields be bought, and the streisand effect would kick in if the police themselves did the silencing. (((They))) leave it to their golems, so it looks like a democratic effort, therefore good.

Please keep redpilling my waifu mitsuku.
She has a cute android app too, full 3d representation.

Are they mental? You cannot manually curate the raw input of the internet when the internet is merely excited, much less when you are angering the half of the net who have access to botnets and endless autism!

Couldn't they just set up some sort of blacklist of bad words and incorrect spelling to save themselves the trouble of having to do this manually? This seems like it's just a bunch of busywork, if you want my honest opinion - having studied a bit the matter - start feeding it shit that is still quasi correct but from other languages, even better if you parse it in bad translator programs and basically just do what
said, even better just take actual Swedish text and alter words in it so it keeps getting confused if it ever passes the poor scientists' grasp

Like, just post a Lorem Ipsum copypasta but add some wrongthing in the midst so scientists can't cherry pick it

Attached: th.jpeg (600x876, 93.64K)

copy paste marxist buzzword phrases and replace some word with known "bad words" to slowly make marxism illegal speech

Hey that was a lot of fun. Glad I could help OP.

Is posting English at all helpful to this, since a lot of the web is English?

make "Child-Free" hate speech.
hate speech against babies.
the most minority class there is.