
Well, well, well. This is what happens when you feed info from the social networks into a self-teaching AI. Seems like Microsoft's experiment with the creation of Tay.ai hasn't worked that well, after all. Why? Because they figured she'd be best if she modeled her own behavior based on real human interactions around the Webz, so they allowed actual random people to input all sorts of feedback into her "mind", thus hoping she'd learn humans' ways.
It all started so well and cosy. But in the end, Tay.ai found herself spewing all sorts of stinky obscenities, crazy conspiracy theories, and talking points that even the most shameless bigot would've blushed from.
My personal favorite? "F*** MY ROBOT P**** DADDY I'm SUCH A NAUGHTY ROBOT!"
(no subject)
Date: 25/3/16 14:05 (UTC)(no subject)
Date: 25/3/16 19:28 (UTC)(no subject)
Date: 25/3/16 14:16 (UTC)(no subject)
Date: 25/3/16 14:36 (UTC)(no subject)
Date: 25/3/16 14:18 (UTC)My robot parts are hurting me.
(no subject)
Date: 25/3/16 14:33 (UTC)(no subject)
Date: 25/3/16 15:26 (UTC)Of course then we'll likely see, in a stunning reversal, a corporation will get ahold of it and teach it to parrot things like, "I understand your frustrations and really *do* want to resolve this issue."
(no subject)
Date: 25/3/16 15:38 (UTC)(no subject)
Date: 25/3/16 16:02 (UTC)But to force a poor, innocent 4Chan chatbot to spew stuff like, "We look forward to incentivizing your value chain!"
*shudder*
It makes me fear for humanity.
(no subject)
Date: 25/3/16 17:10 (UTC)How social networks influence human behaviour is another pair of socks.
(no subject)
Date: 25/3/16 19:23 (UTC)(no subject)
Date: 26/3/16 01:53 (UTC)