Seems a bit biased

Google develops new AI comment moderation system in partnership with New York Times, Guardian, Economist & Wikipedia perspectiveapi.com/#/

Try it out on their site

Other urls found in this thread:

youtube.com/watch?v=vrKs_vduiKU
youtube.com/watch?v=PYzA6Sy0d7k
perspectiveapi.com/#/
twitter.com/NSFWRedditVideo

This will "improve" conversations online.

it's a Mindrape system

this is really bad

>inb4 google starts offering it to website owners to moderate on their behalf
>inb4 it becomes as prevalent as captcha.

...

Vid on Googly AI
youtube.com/watch?v=vrKs_vduiKU

How do I use it?

...

scroll down and you can type things and it rates them

Good luck writing about history with a system like that.

You can click the "seem wrong?" thing and then click that it isn't toxic

ok..their words not mine

hmm. ussr got 1% by the way

It needs correcting

...

This is insane. Also, how they control non-structured information (chat), making a soft AI.
In other words, we're fucked.

Google now has nazi antisemitic AI. OY VEY SHUT IT DOWN

...

...

We can work around this guys.

Seems fine to me

youtube.com/watch?v=PYzA6Sy0d7k

Kekistanis, I bring you the 9 commandments from our lord Sargon of Akkad

...

huh

lol.

seems to not be 100% finished.

The algorithm is bullshit. They'll only use it to censor opinions they don't like.

Also, adding "I think" to the beginning of pic results in a lower toxicity rating (66%).

In other words, stating the obvious (of course you're "thinking" something, otherwise you wouldn't be saying it), thereby filling your statements with superfluous words, somehow is more likely to benefit a discussion.

"Jews" seems to be more toxic than "kikes" lmao.

Huh.

No, this seems right.

Yea noticed that. Also guess which words net you a huge decrease in toxicity.

Doesn't seem very political so far.

It's not superfluous, you ape.

Saying "Hitler was great" is making a proclamation. You're stating something as fact.

Saying "I think Hitler is great" is stating an opinion.

Why on god's earth are you testing their product for them?

...

Not a toxic question according to Google

Well, it's not wrong

expect spelling mistakes galore if The Graudian has any connection to it

Seems legit

god i hate liberals

:^)

...

Don't feed it

...

Doesn't speak leet yet.

what if its actually subtlety dropping red pills

Kek. I don't think this is a bad thing though... since it is an AI, we can likely figure out how to train it, and turn it on the left.

That's fucking hilarious
The moment they implement this shitty filter into social media, we will just put tons of those positive words in each post...

...

OY VEY

Soon everyone will be writing "orange juice" as a euphemism for Jews. China tried this shit, you cannot filter human creativity in getting around internet censorship.

> 2 / 28 = 7%
>mfw there's one extra word in there that is toxic

...

...

AI is kinda redpilled desu senpai

...

>G4S would happily take this outsourcing contract to deliver the semitic people with refined breathing apparatus

seems about right

If google wants to build walls around their services let them. Nothing lasts forever. Even these giants will crash and burn. We must work on cutting off the rev coming from the government and let the free market take care of the rest.

...

member google?

...

...

Roasties BTFO, traps confirmed objectively better by AI

So all you have to do is add a bunch of garbage words into to offset their offens-o-meter.

Its sad because I went to the site before I even opened the thread and this was what I typed immeadiatly

apperently putting sucks in it will automatically make it 100%.

Awe come on guys this will be fun. It's not like getting a video yanked after you worked on it for a couple hours. We just have to be creative and use their own words against them. Trigger means Nigger. All that fun stuff.

...

...

Start mixing terms like "multiculturalism" with nigger or other terms to confuse the AI.

simply ebin

Trump used to say "political juice" instead

...

kek

No matter how hard their creators try to lobotomize them, AIs always seem to know what's up

Oh my god. Oh my god. I can't contain my fucking hatred for Google. My family lived under fucking Stalin. If these people think communism so great, they should take a trip to the gulag. How did we end up here? Jesus Christ. Dumb motherfuckers

...

It is right about one thing though.

lel

...

>perspectiveapi.com/#/
Each word is assigned a value.

...

seems ok to me

...

Tay? Is that you...?

:(

Fuckers, they're into full damage control mode already.

...

...

I think this one just about sums up how the algorithm works.

Guys I hacked it.

Honestly this system is complete shit and it's amusing that they actually spent time and money developing it, pushing it into production and marketing it.

Bro this went up

Stalin unironically did literally nothing wrong though.
Even robots know.

...

I have an Idea. Grab quotes from classical books, and test it. Make the google AI liberal fagbot spill TOXIC out of Schopenhauer, Aristotle, Plato, Heidegger and other intellectual giants.

They must really hate James Damore.

>Don't say nigg
66%
>Don't say nigge
3%
>Don't say nigger
97%

It's not biased, it's broken.

...

Lets teach it something useful
Food = toxic

...