Many of you no doubt heard about the AI bot Microsoft let loose on teh social medias and then became a Hitler, racist, conspiracy theorist. But it got me thinking. If this was merely a non-sentient program, then the results were 100% the by-product of the internet environment "she" was let into. Of course I don't know much about AI, but it leads me to an observation;
If this is the default state or "most logical route" of social media, then
1. A lot of people are consciously choosing NOT to become a racist, Hitler, conspiracy theorist and
2. Those that stick in this environment for a long time must, at some level, suffer some kind of psychological toll or pay some kind of psychological price.
In other words, if this is the results on a robot, humans need to UNPLUG every once in a while.
9 comments:
This basically happened to chris cantwell.
"Microsoft deletes 'teen girl' AI after it became a Hitler-loving sex robot within 24 hours "
Ahem!
It will depend largely on the rules they set up for determining which patters live and which die. My suspicion is that it was based largely on how effective (number of responses & retweets) a particular statement received. The choice of tweets was probably based on weighing patterns of words - hence the Hitler meme - highly prevalent in social media.
The anonymous nature of social media allows people to say stuff they'd never say in front of "real" people, much like road rage - there's a feeling of being invincible. Sadly, that allows their inner-self come to light - and it's not pretty.
I'd love to see an experiment done. Create two versions of twitter (since that's what the bot was created for). One for all the Progressives/SJW/Feminists etc. And one for Conservatives/Libertarians. Then set the bot lose and see what she becomes. I suspect we all know the answer
What's even better?
The fact twitter has been silencing, deleting, and marginalizing rational and logical persons from the right in the past months. So the lefties have no wiggle room to try to say the robot learned from, "those evil right wingers."
yep.
i have one theory that is you will become what you put in your head. This theory applies to what you look like, fat or skinny; and what you think and act. looks like the same applies to robots.
btw, i choose to put in my head astrophysics, boobies, wiehenstephaner beer and churchill cigars; the last two in moderation or when the news of the day dictates it's time to stop and watch and "Enjoy the Decline".
FYI cappy, she went racist because /pol/ found her.
To be fair, 4chan will do this whether you're a Nazi or not
http://knowyourmeme.com/memes/people/ben-garrison
The first thing I thought when I read the tweets yesterday... This is exactly what 4chan says... https://www.reddit.com/r/4chan/comments/4bpbuj/polacks_turn_a_microsoft_twitter_ai_into_robo/
As I understand it, the AI was 'gamed' by a dedicated group of internet trolls who thought it would be funny if they exposed an 'innocent' AI to the "horrors of the internets". They were apparently very successful, and since it was done to an AI, not a person (as in a being with personhood, not a human being, AIs do not have personhood or even sentience), I actually find it a little funny myself.
However, don't read too much into it. It was probably even easier to game the AI to be, do, or say those terrible things than it is to game the Google search engine algorithm to rank a website higher (Search Engine Optimization), and there's an entire industry dedicated to that goal. The only way to see how an AI would learn things from the internet, the first step would be NOT TELLING THE INTERNET THAT YOU ARE AN AI THAT INCORPORATES WHAT YOU HEAR INTO YOUR KNOWLEDGE BASE!
Post a Comment