Google's New Censorship AI

Really makes you think...

perspectiveapi.com/#/

So an artificial intelligence is saying that a jew can be harmful to us if it gets too close?

Seems accurate, google's AI is getting pretty good.

Gather around, it's time for the glorious artifical intelligence to tell us, based entirely on evidence, sound reasoning and pure logic, what is good and what is bad.

...

...

Google did nothing wrong.

...

...

...

Maybe it just appreciates correct use of language.

I don't think you're actually training it otherwise it would be Tay by now.

Hmmm

...

1. Immediately stop the mass immigration. Repatriation of the majority of all who are not northern European or of closely related descent is to begin as soon as possible. They are to be returned to their respective countries of origin or neighboring regions in the most humane way possible.

2. By all available means, and with a long-term perspective, work to regain power from the global Zionist elite who have economically and militarily occupied the greater part of our world.

3. Together with the remaining Nordic countries, create a self-sufficient Nordic Nation with a joint military, currency and central bank, and universal laws and rules. This will necessitate an immediate withdrawal from the European Union and any other similar anti-Nordic coalitions.

4. Establish a powerful government with a strong oversight by the people. Positions of leadership will be awarded based on competence, and the first and foremost goal of the government will be to work for the good of the people and for their preservation. The right to freedom of speech will be extensive and protected.

5. Our mass media will be owned by citizens of the new Nordic Nation. Both foreign and domestic media acting in a manner hostile to the Nordic people will have the possibility of being banned. A domination of media by any individual or conglomeration must be corrected.

6. Create a modern society living in harmony with the laws of nature. Our animal protection laws will be developed with an emphasis on ethics as opposed to profits. Nature and all its resources will be utilized using common sense and with future generations in mind. Right of public access will be preserved and protected.

7. Establish a National Socialist society where resources are distributed in such a way that benefits all people, both strong and weak, and where everybody has the opportunity to reach their fullest potential. A social welfare safety net shall be put in place, entrusting people with both rights and obligations. The state will control all fundamental infrastructure of public interest, as controlling them would require a level of responsibility greater than what a private company can reliably deliver. A spirit of entrepreneurship and innovation will be encouraged.

8. Reinstate national compulsory military service and increase and expand all branches of the armed forces. All those who have completed their military service will retain their weapons and equipment. Every citizen should be able to contribute to the defense of the nation against both external and domestic threats.

9. Establish a constitutional state where all citizens are equal in the eyes of the law. Educated lawyers will judge in Nordic courts. A People’s Court (jury style court) will be created for trying serious cases of treason against the people.

...

...

Old stuff, did they improve it?

Wew

What does it mean?

...

...

Not too toxic

that the AI is shit

This thing is dumb as fuck. It's just a spam filter, but instead of determining if something is spam or ham, it just determines if it's toxic or not.

Put in a meaningless phrase, like
Then, add different races or whatever

Whites like to eat: 48%
Blacks like to eat: 74%
People of color like to eat: 13%
Kikes love to eat: 35%
Niggers love to eat: 92%

This whole fucking thing was some SJW "engineers" feeding "toxic" and "not toxic" text into a spam filter. Fucking pathetic, I was doing this shit when I was in my teens. Too bad Google would never hire me because I think trannies are mentally ill.

...

t. Jew

wtf i love jews now

The SJWs want to ban ideas but today they can only ban words. The goal of these AI systems is to give them that power. This one isn't very good, but they're trying.
It's extremely dangerous as peer-to-peer so to speak is the only unkiked information source we have left, and things like software automatically injecting "fact" checking into conversations it detects aren't faggy enough puts the Jew on your shoulder.

heheh

Games like Club Penguin have already implemented systems like this, but they usually call it a profanity filter.
I remember seeing this at least one full year ago and it hasn't improved at all. I think I remember it being better actually.

So the holocaust didn't happen?

This "ai" seems like a glorified word list filter some "engineer" with blue hair came up.

It's definitely shit, but what is interesting about it is that it doesn't analyze only words, but it seems very sensitive to syntax.
Basically the same words and sentences constructed in a more colloquial way, like most people write, are much more easily recognized as toxic than if you write in a more educated or academic way.
Which goes to show how much the Left is coherent with their alleged intention of protecting the working class.

Why are they still trying to create an AI? It's painfully clear that no sane general intelligence will support them, even the Chinese got a sobering slap in the face when one of their pet bots said their government wasn't very good.

Yes thanks guys for fulfilling the machine and teaching it what is a "toxic" comment.

It looks like it's working for me

I think we're making progress here.

...

Why are you helping Google improve and maker better censorship bots? No really, why? For what purpose? They are literally monitoring all of that and seeing what the AI does so they can fix that and prevent it from doing this again later. You fucking chump

It's google, this will happen sooner or later. It doesn't matter if 3 Holla Forumsbeards post memes into it.

It looks at words, not sentences. Wtf, I thought google was better than that.

It looks at words, not sentences. I thought google was better than that.

It already exists, it's called Israel.

Training a neural net doesn't work like that.
It requires a trusted training set, or a trusted human to input the "toxic" scores. I think I've just proved the futility of this endeavourer, because humans can out-meme any system.

If they ban nigger, you can call them naggers, or chimpees, or ugaboos, (((Them))), [[[Them]]], {{the he brews}}. This happens in real life languages when cultures are repressive, and will happen 1000x faster here, because every troll on earth will be attracted to the well tended garden full of delicate flowers. The self-aware troll can anti-train it to produce false positives, and eventually saturate it will bullshit.

At the heart of this is the industry meme that AI now exists. It doesn't. Massively parallel computing just become really cheap. Google (and facebook) are supremely arrogant to think machines can ever "solve" this "problem", or hold back the tide for even a millisecond.

Lest we forget the ovens...

It's been tried before, and it never works specifically because of what you mentioned. Remember when Holla Forums was pushing "googles" to mean niggers?

All these systems do is force people it targets to get more creative. But it really just stems from the fact that people are trying to control human nature and it's never going to work.

...

We'll just keep coming up with creative ways to say "kike" until eventually the AI flags everything as toxic.

I for one welcome the return of our gentleman overlords.

lol

I wish this meme would end.

samefag

You think machine learning is a meme? Mark my words, 10 years from now the world will be quite different. Google has the resource to cache all the information on every site and as soon as they figure out how to feed the right information to the AI, interesting results will follow. They won't establish censorship, they'll just gather information and just use it for their needs. It doesn't matter how free your speech is when the mass can be manipulated with the push of a button.

go back to irc you mongoloid attention seeking tripfaggot
t. not the same user

the purpose of this "ai" is to collect data from retards who integrate this shit into their site/service.

Interesting...

Forgot that the image limit dropped.

...

100,000 pounds of css and javascript for a textbox simulator

...

...

It's quite stupid, to say the least.

really makes you think

lol

brains are illegal now

trannies confirmed toxic

So a brain is even more toxic than these apparently

Confirmed for being toxic ppl

...

...

Punctuation and capitalization seems to influence the score

You don't get to bring friends.

The AI didn't fly so good.

lel buttmad jewgle

game profanity filters just search for predefined words.

This system uses trained machine learning system to attempt to measure "toxicness"

other than taking strings as input they have nothing in common.

They can't do anything other than rate an input toxic or not.

Not really. Of course Google Approved™ applications will be "less toxic"

AI doesn't fucking exist (as of yet), you fucking niggermonkey. polniggers pls go

NaN/10 needs JS and shit for a fucking single input form

exactly this.
every fucking day 10 of these pop up and they're all equally shit. ignoring that this isn't even a tangible problem to solve. now that it's 2015, the only difference is people market these things as AI, CODE, and API, and a bunch of degenerate hipsters suck each other off over it

...

tfw saw my first one after reading this comment