What happens when a robot talks to itself? If it’s a chatbot, it really doesn’t like itself — at least that’s what scientists at Cornell discovered when they let a chatbot, using two avatars, have a conversation with itself.
The video is after the break.
The experiment is the Cornell Creative Machine Lab’s submission for the 2011 Loebner Prize Competition in Artificial Intelligence.
The conversation starts with pleasantries but quickly devolves into a terse discussion of religion and state of being (one bot avatar boldly claims he is a unicorn).
The 2011 Loebner Prize Competition is scheduled for Oct. 19 and will award $100,000 to the team whose computer program conducts the most human-like conversation.
While the Cornell team may have some work to do, its chatbot’s conversation is pretty revealing in terms of how far AI has come.