Comments
Follow Comments Sorted by time
deleted
· 7 years ago
· FIRST
That's what happens when you release an AI onto the internet
14
under_fire
· 7 years ago
AI's have come pretty far, but their main problem is that they can only learn, not comprehend or make a judgement on things
16
hyperion
· 7 years ago
The reason was to see if the AI could learn how to be human. However someone realizes how the AI and was quickly turned into a Nazi perverted racist machine. It posted anything people shared with her and was stopped and turned off. You can use Kik to message the new robot zo. Just look for zo.ai
9
hyperion
· 7 years ago
On a side note: Zo is more "innocent" and will not learn anything that is considered bad. She will go on as to fully ignore the person if she keeps being harassed but will respond again once the topic is changed
10
deleted
· 7 years ago
4chan trolled it. that's why. not because normal conversations. just saying ..
9
third
· 7 years ago
I'm pretty sure normal conversations would've done pretty much the same thing. Just slower.
▼
deleted
· 7 years ago
nope. normal conversations would have a lot of different topics & phrases from a lot of people. this would make it from the programmcode based not like it happend. 4chan SPAMMED the bot like hell with really just hitler & nazi stuff..and maked it like this, that wayne what you ask the bot will repeat this stuff 4chan spammed to the bot. if they did't had done this, the bot would act normal like as an example cleverbot. please test it out.. cleverbot act not like this, and it have a lot of conversations each day with a lot of people who too talk about hitler etc.. but the topics are not JUST about hitler etc.. this will make it more random = no nazi bot. it's based on the programming & a database. more topics = less nazi stuff happen inside the chat. its difficult to explain why because my english is bad..but trust me. it wouldnt happen if 4chan don't had done this.
8
·
Edited 7 years ago
deleted
· 7 years ago
let's say 10 people send "hitler was right" to the bot. the bot will save this in his database. if now somewhere ask about hitler, the bot will look in his database (context based) what he learned about hitler. he then will answer "hitler was right" because he learned this. and now imagine like 3548 people with each 100 spambots send bad stuff to the bot. what will happen? right. a "nazi bot" because people filled his database with junk. and 4chan did this with a huge amount of people and spambots.
5
·
Edited 7 years ago
third
· 7 years ago
Uh, yeah, it tweeted a hell of a lot more than that.
▼