Google develops new AI comment moderation system in partnership with New York Times, Guardian, Economist & Wikipedia perspectiveapi.com
Try it out on their site
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
youtube.com
perspectiveapi.com
twitter.com
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
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
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
...