Microsoft launches Twitter Chatbot which immediately learns to be racist

Status
Not open for further replies.
In a lot of ways this is a hilarious microcosm of any white millennial twitter user, we're programmed to respond in certain ways to conversation and depending on our influences coming in it wouldn't take much to push some people into saying the stupidest things just to fit in with a crowd.

In other ways it's not funny because a lot of these people trying to influence the bot actually believe the shit they're saying. :(

I'll just keep in firmly in the funny camp because this kind of twisted Black Mirror-esque story doesn't come around every day.
 
Any recorded archive of this all? It seems Microsoft purged and reset the chatbot.

FmYzlPf.gif
 
I give Microsoft credit for the balls to make what seems like a true unfiltered AI. I mean fuck, there is no language filter and no logic on what is offensive or not. It's like making Siri dictate offensive shit in AI form
 
If I had to guess, this bot follows people who follow it as well as the people that they follow. It then scours through the tweets to find posting patterns.

It's like a reverse gg-autoblocker only without the filtering that gg-autoblocker provides to make Twitter an actual usable space.
 
If I had to guess, this bot follows people who follow it as well as the people that they follow. It then scours through the tweets to find posting patterns.

It's like a reverse gg-autoblocker only without the filtering that gg-autoblocker provides to make Twitter an actual usable space.

Also that is probably why it will only get the worst of people. As soon as blockers started...blocking it it no longer had access to what those people were saying to add to itself.
 
I assume they did.

It's just mirroring the internet back at itself though. So honestly it reflects more poorly on the people using it than MS.

This. Interesting stuff. Like a mirror of the Twitter society. As someone else said it's a black mirror episode.
 
This both hilarious and highly interesting. This isn't exactly the same but there is an article about how IBM wanted their AI program/system Watson to learn to converse in a more natural way by using slang. So they they of course uploaded the contents of Urban Dictionary to its memory and it got foul mouthed very quickly since it didn't really filter out what was profanity and what was fit for common conversation.

http://fortune.com/2013/01/07/teaching-ibms-watson-the-meaning-of-omg/

It even responded to a researchers question with the answer,"bull shit". My kind of AI.
 
Status
Not open for further replies.
Top Bottom