Here is some more from the European Computing and Philosophy Conference ECAP 2007.
The session “Computational and post-computational approaches to the mind” was chaired by Susan Stuart. (I will learn more about Susan tomorrow in our joint panel session on the “Future of Artificial Intelligence”).
From the two talks I found the presentation “Turing test: mindless game? – A reflection on the Loebner prize” by Huma Shah and Kevin Warwick quite enlightening. I have to admit that I did not even know that there is a yearly contest. But the format of the original Turing Test has been altered considerably. Whereas Turing suggested an amount of five minutes for judging if one talks to (i.e., text-messaging with) a machine or human, the Loebner prize format has 20-minutes time-slots. And there are other alterations that change the Turing Test considerably. Let me just point out that the authors still believe that the Turing Test is useful, but the Loebner prize is deemed problematic at least. Positive observations comparing Loebner prizes in 2001 and 2006 are that conversations now can take place at length. But the negative aspects outweigh this because still “no meaning [is] transmitted or understood in conversational transactions”. Still keyword spotting as in Joseph Weizenbaum’s Eliza is predominant. The answers are still non-contextual.
[composed and posted with ecto]