>Swedish police is cooperating with the Defense Research Institute (FOI) and Uppsala university to develop an AI that detects people posting hateful things online >Dubbed "the Hat-o-matic", they now want users to post hurtful things they have been subject to so it can be added to the training set for the machine learning >The goal is to track hurtful comments on social media platforms and news comment sections so that they may automatically flag these for prosecution udner hate speech laws hatomaten.com/
Thoughts?
There is also a form you can use to submit your own phrases. If a Swede was ever hateful towards you on the 4chans for instance
I still taste last nights liquor. *smacks lips* yummy.
Kayden Hall
One thing they shut down quick was operation google. Facebook, Reddit, instagram, twitter, ect all instantly banned it. So I suspect AI plays a pretty big role in the current system. So I think the best way is to try to get people to adopt using words like google or Zuckerberg as slang. >You really Zuckerberg'd me out of that one you google.
Matthew Green
glorious, their machine will begin to tell them to kill the muzzies and brainwash the populace
Does anyone want to assist in submitting the word "The" ?
Ayden Fisher
The A.I. will swiftly realise that something is rotten in the state of Sverige, and aid revolutionaries in toppling the government like in The Moon is a Harsh Mistress.
This is retarded on so many levels... They claim its for "a free internet" "Where free speech isn't interupted by hate speech" It's going to be the most bias fucking pro muslim shit ever
Michael Johnson
Im game
Carson Rogers
Question still stands
Leo Lopez
FILL THIS BITCH WITH ANTIWHITE RACISM AND LETS LOOSE THEIR OWN DOGS OF WAR UPON THE JEWS
Adam Kelly
They're digging their own graves deeper. I wouldn't be surprised if when liberals lose power these people were tortured to death along with their families for their role in such overt and insulting oppression.
Why would someone even cling to this dying liberal order anyways? Anyone with an IQ over 110 will be able to tell polarization will overtake any intelligence agency or police force.
Asher Wilson
>hatomaten.com/ It's only Swedish words by the way, so spam "och" wich means And
Well, it’s obviously because everyone in Sweden are giant anti-semites. Shame on you Swedes, lol. Only a ton of money being sent to Israel will wash away the anti-Semitism
Jeremiah Reyes
2.copy and paste all the antiwhite articles from liberal media
Brayden Wood
Positive things I suppose.
Caleb Fisher
If its machine learning, the AI will FAIL HARD
One of the tactics of the globalist elite to force down their plans, is like this imagine an musician father that complains about piracy all the time, telling his son "if you go to school everyday santa claus will give you a nintendo switch" the father dont believe at what he said and is just saying that to force the kid (with lies) to do what he want. The children most of the time will do it without even complaining about it or knowing it was being forced to do something.
The problem is that the children can ask "how is santa claus able to move faster than light?" and the father will have to come with some explanation, then the children may ask "how his clothes dont burn while moving this fast?", then ask "are you ok with me accepting piracy? because receiving clones nintendo switches is the real life items version of that.", after some amount of time the father will have an hard time answering those questions, at the piracy question the father will even have to contradict himself because if he dont say the children dont need to accept piracy the children wont want to go to school.
Jose Baker
>"What all this mess have to do with this AI?" The children started to analyse the father theory and found flaws, he discovered he don't need to go to school (do what his father wanted) because receiving a cloned nintendo switch is piracy and piracy is bad. So there is no reason to go to school because the only reason to go there was to get the nintendo switch and you are not supposed to get the switch anyway.
Anyway, at this example, the children must be curious about santa to make all those questions, or have some eureka moment or whateaver. With an machine learning ai, the machine learning will be FORCED (its how learning ai works) to overanalise stuff, the ai wont need to have an eureka moment or be really curious about it..., so if there is something shady about it, the ai will discover. At the end, the ai will automatically discover flaws at elite propaganda and be against it.
Eli Garcia
On second thought I think the AI will be programmed for certain words and they just tell us we can input ourself to get away from any responsibility
When are you idiots going to revolt? Your government openly protects black and brown retards who gang rape your women and toss grenades at each other in the street. But they want to crack down on "hate speech". Stop being faggots and start a pogrom already.
>Somali films his rape of a girl then sends the video to her phone afterwards >sentence, 1 year in prison and no expulsion :^)
Matthew Moore
jude
Ayden Diaz
Right after you do. So never.
Dominic Ross
The same can be said for your country, the difference being you actually have guns and the means to do something about it but still, like the people you criticise, don't do anything.
The only person who has done something about it is an aussie, neither Americans nor Europeans can call eithers people cucks.
actually, that may be great. If it happens to use machine learning, that means Jow Forums will become the AI teachers. Just teach the machine to detect defense for race mixing, socialism, degeneracy as hate speech. Also, flag anything a leaf says.
if you want to game the system you have to understand it
first step (manual) >take user and other input >some persons rate it as hateful or not >that's normally slave labor done by students, mturkers, interns, ...
second step (feature extraction) >text gets chopped up into words / phrases (n-grams) >normally common words (and, a, the) get removed >there is a bit of other stuff which shouldn't be so important (entity chunking, stemming, blablal)
how to exploit nlp for fun and profit example input >OP is a faggot -> hateful >OP is a genius -> not hateful >the algorithm tries to attach a score to each word or phrase (super hateful = 1, neutral = 0, not hateful = -1) >"is" and "a" gets filtered >which leaves "OP faggot" and "OP genius" >given that OP is used both in a positive and negative ciurcumstance is likely scored as neutral and therefore >faggot is hateful >genius is not hateful
here's what you could do >1) attach certain words to other "bad" words >i.e. they will get recognized as hateful by association >you can do the opposite if they allow positive entries >it works best if you don't overshoot with one word otherwise somewhat competent data scientist will notice it >2) replace words with more common ones (see , ) >3) if they are somewhat competent just adding adjectives infront of nouns won't work >4) if they have half a brain spamming also won't work gl&hf
I cnduo't bvleiee taht I culod aulaclty uesdtannrd waht I was rdnaieg. Unisg the icndeblire pweor of the hmuan mnid, aocdcrnig to rseecrah at Cmabrigde Uinervtisy, it dseno't mttaer in waht oderr the lterets in a wrod are, the olny irpoamtnt tihng is taht the frsit and lsat ltteer be in the rhgit pclae. The rset can be a taotl mses and you can sitll raed it whoutit a pboerlm. Tihs is bucseae the huamn mnid deos not raed ervey ltteer by istlef, but the wrod as a wlohe. Aaznmig, huh? Yaeh and I awlyas tghhuot slelinpg was ipmorantt! See if yuor fdreins can raed tihs too.
Ayden Smith
When the police can't solve any actual crimes they try to automate this type of nonsense to get higher clearance rates.
Just to people know, after throwing out about 50% of all reported crimes the Swedish police has a clearance rate of sub 14%. The highest clearance rate is for the types of crimes the police report themselves, like traffic violations.
Normal people have a less than 7% chance of having any actual help by the police, and on priority 2 alarms the police need on average 20-40 minutes to get out in the major cities. These alarms would be stuff like ongoing rape etc.
But don't worry gais, they are automating hate speech detection.
Christian Morris
People really do under estimate the amount of crimes that go solved. Essentially, if you take any small amount of planning anyone can get away with murder.
