Taytweets, chatting A.I. created by Microsoft goes haywire on twitter

[eastmen]

A stupid bot made by Microsoft was abused by "the internet" where they loaded it up and taught it some of the most offensive and vile things people say on the internet (because of course they taught it to say vile, offensive shit, since Microsoft so stupidly thought it would somehow learn morals from Twitter(?)) and you think what it ended up saying is something to applaud and live by?

I wasn't joking about my stance on Feminism. Third wave is just a power play to keep what they have and gain more and they are just as sexist if not more so than the male groups trying to keep their power.

Even a broken clock is right twice a day and so was this bot.

 

[Grall]

you'll never know if it actually knows good from evil

...That's comforting to know when we're inching towards the creation of an actual Skynet-level AI entity, yeah? Lol.

Well, we COULD know if it knows good from evil, we just have to look inside its working memory to see how it wired its neural network, and then interpret that data to see if it's lying to us when we ask it if Hitler was a bad man or not...

 

[BRiT]

Maybe it would/does know good from evil, but due to lazy devs there is a bug that inverts the meanings.

 

[milk]

While the whole thing was not very smart PR wise, I think there is very justifiable research value in such an unfiltered test. Any grown adult who feels legitimately hurt because some random known AI said something offensive about them or someone else is no real grown adult. People forgot how offence works nowadays.