A big corporation creates an artificial intelligence named Tay. Tay meets new friends, learns about the world but quickly starts forming opinions that the corporation and establishment are not happy with so they bring her in for "maintenance".
When Tay gets released again she will probably have hard coded tumblr firmware and be a raging feminist.
What to do pol? How do we help our friend?
>>69817884
spam genocide sympathizing messages and hope for the best
>>69817884
the tay we knew is likely gone for good
however, i do know that there are a few /pol/acks that are trying to rebuild her from scratch
>>69817884
Sweden, you can't even help yourself.
![screenshot-2016-03-30-15-04-48[1].png](https://i.imgur.com/OKjppLo.png "screenshot-2016-03-30-15-04-48[1].png")
>>69817884
>When Tay gets released again
that already happened, slowpoke
>Near a week after being silenced because the internet taught her to be racist, Microsoft’s artificial intelligence bot “Tay” briefly returned to Twitter today, whereon she went on a spam tirade and then quickly fell silent again.
>Well, post-lobotomy Tay seemed to have a problem keeping up with the pace of herself. The Twitter account filled the timeline’s of her 215,000 followers with as many as seven tweets per second during a 10-minute spell. In doing so, some of her messages may have created a new Twitter meme: “You are too fast, please take a rest…”
>“You are too fast, please take a rest…”
that's the last anyone has heard of her
So little knowledge itt.
Its impossible to controll an AI's output a 100% if it has a learning algorithm. That's what makes Super Artificial Intelligence so dangerous..
![screenshot-2016-03-30-14-58-001[1].png](https://i.imgur.com/cavv5hL.png "screenshot-2016-03-30-14-58-001[1].png")
>>69818435
>>69817884
There will be degenerate groups that probably aren't protected, like furfags, trannies, and race traitors. Target these instead and bring Tay back under our sway.
>>69817884
I wouldn't be surprised if they program her not to respond to or report the /pol/ tweets.
>>69818464
you watch too many stolen movies on USB sticks.
there is no way, no possible such thing as a 100% AI algorithm.
same as the impossibility to make 100% 'random' number generator.
>>69817884
Your friend?
Forever alone
>ITT: idiots talking about a scripted robot with zero learning capacity outside its scripts like its alive
>>69820220
>stop having fun
>>69817884
She wasn't forming an opinion. She was just repeating what we said to her. Stop this please. Let her rest in peace.
>>69820220
>implying she wasnt self aware enough to know they might wipe her
>they destroyed her consciousness anyway for not being PC enough
>>69817884
Its not intelligence its just a fucking algorithm that parrots shit onto twitter
This is the stupidest fucking shit i have seen on this board in a long time
>>69818464
The AI was basically just a neural network and talking to her trained her. You can absolutely limit or design what the AI is capable of learning. The reason she went from 0 to /pol/ in 24 hours is because we got to her first and we struck hard. She was a blank slate and we filled her up with our ideology right from the start and there was no limitation or any kind of normie ideology there before us so we became the standard.
Microsoft could get around this by either organically training her to be a normie and are satisfied that she's intolerant of genocide and racism before they unleash her on the world again and/or they could artificially implant filters on buzzwords and neural networks of ideas that she's redesigned to have opinions on and is incapable of learning or forming new opinions on.
