I did some research on this a while ago. Can they learn from previous conversations and parse large corpuses of text? Yep. Is there an open-source project? Ye

I did some research on this a while ago. Can they learn from previous conversations and parse large corpuses of text? Yep. Is there an open-source project? Yes - https://en.wikipedia.org/wiki/Artificial_Linguistic_Internet_Computer_Entity https://en.wikipedia.org/wiki/Artificial_Linguistic_Internet... . It's, AFAICT, somewhat behind the state of the art ( http://www.loebner.net/Prizef/loebner-prize.html http://www.loebner.net/Prizef/loebner-prize.html ), but still very interesting.