rosencrantz Posted March 24, 2016 Share Posted March 24, 2016 Look what twitter did to Microsoft's Tay https://twitter.com/geraldmellor/status/712880710328139776 Link to comment Share on other sites More sharing options...
AccuTron Posted March 24, 2016 Share Posted March 24, 2016 I had trouble figuring out what was even going on. What's going on? Link to comment Share on other sites More sharing options...
AncapFTW Posted March 24, 2016 Share Posted March 24, 2016 I had trouble figuring out what was even going on. What's going on? No idea, though the top tweet by the guy who you linked to was "seems like the right to bear arms ought to be outweighed by the right to go to school and come home alive." Apparently that's relevant somehow? Also, how is that even a coherent statement? Link to comment Share on other sites More sharing options...
shirgall Posted March 24, 2016 Share Posted March 24, 2016 Full disclosure: I work for Microsoft, but not on this project. Microsoft put out an AI bot on twitter that would communicate with millennials similar to other chat bots in the past like ELISA would do. It was intended to learn from its conversations and synthesize new and interesting things to say by lifting elements from replies and combining them with some pre-existing rules. Some vandals came along and seeded the AI with a lot of racist, violent, and otherwise ugly things to say and eventually got it to start replying to people with distasteful things to say. Gone are the days when you can put out something cool or interesting and people would be polite enough not to subvert it immediately. Link to comment Share on other sites More sharing options...
AncapFTW Posted March 25, 2016 Share Posted March 25, 2016 Full disclosure: I work for Microsoft, but not on this project. Microsoft put out an AI bot on twitter that would communicate with millennials similar to other chat bots in the past like ELISA would do. It was intended to learn from its conversations and synthesize new and interesting things to say by lifting elements from replies and combining them with some pre-existing rules. Some vandals came along and seeded the AI with a lot of racist, violent, and otherwise ugly things to say and eventually got it to start replying to people with distasteful things to say. Gone are the days when you can put out something cool or interesting and people would be polite enough not to subvert it immediately. Reminds me of chatterbots I used to play around with in high school. Most of the ones on the website would cuss at you, say racist/sexist things, etc. I tried to get a new one to act like a butler, since the others were too far gone, but I didn't do enough with it. Link to comment Share on other sites More sharing options...
Crallask Posted March 25, 2016 Share Posted March 25, 2016 It did come up with some genuinely hilarious responses though. Link to comment Share on other sites More sharing options...
vahleeb Posted April 2, 2016 Share Posted April 2, 2016 https://youtu.be/Tv74KIs8I7A?t=13m46s This video has a pretty nice handle on what happened. Link to comment Share on other sites More sharing options...
Recommended Posts