Thursday, July 28, 2016
Artificial Intelligence is not ready for twitter trolls
With the victory of Alphago in the ancient board game of Go against the world champion, Lee Se-dol, many people became worry that Artificial Intelligence is getting way to smart that it can someday outperform us in daily objectives. However they cannot outperform us in using social media websites such as Twitter or Facebook as demonstrated by Microsoft’s failed AI Tay.ai who became very racist after people took advantage of Tay’s learning capabilities. For sure this was a sign of relief for people who were worried that a computer can someday take their job and for others was a funny prank that offended many people. What I found very fascinating about Tay was her ability of being able to respond to people with a immediate response proving that AIs might have better chance of communicating with others, but does not have a filter that blocks any type of offensive language or hate. What do you guys think about this AI? Was she a success or a failure?
Subscribe to:
Post Comments (Atom)
This iteration of an Artificial Intelligence is simply a step toward something more powerful. It's highly likely that as soon as that attempt went bad, the developers went back and tried to fix things. It's all progress. And the only thing that will hinder this progress is the people working on these projects giving up or deciding that they shouldn't be doing what they are doing, and neither of those things have happened universally (if at all) yet.
ReplyDeleteThis iteration of an Artificial Intelligence is simply a step toward something more powerful. It's highly likely that as soon as that attempt went bad, the developers went back and tried to fix things. It's all progress. And the only thing that will hinder this progress is the people working on these projects giving up or deciding that they shouldn't be doing what they are doing, and neither of those things have happened universally (if at all) yet.
ReplyDelete