Fifth Talk in the Alan Turing Year at Reykjavík University

The fifth talk in the Alan Turing Year at Reykjavík University was delivered this afternoon by my colleagues Yngvi Björnsson and Kristinn R. Thórisson. The talk was entitled Alan Turing's Contributions to Artificial Intelligence: Can Machines Think? and has been organized in collaboration with CADIA and IIIM. This was a thought-provoking and very enjoyable scientific event. In case you are interested the audio and the slides of the talk are here in .avi format. (Note: For technical reasons only the audio of Kristinn's presentation is available.)

In his presentation, Yngvi introduced the field of AI, its subbranches (applied AI, strong AI and cognitive AI) and highlighted Turing's main contributions to the field. On the other hand, Kristinn presented a critique of the Turing Test. Kristinn is a firm supporter of strong AI and his position on this matter can be summarized as follows. (I hope that I am not misrepresenting his views too much.)

1. The standard divide-and-conquer approach that we use in science to understand phenomena is not going to help us understand "intelligence", at least not if applied in the same way as has been done so far in AI, namely by using it in a reductionist way to remove features that are central to the phenomenon of intelligence.
2. The Turing Test was a very premature attempt at devising a test for the phenomenon of intelligence that forced upon much constructionist AI research the view that "intelligence is X, where X is some very simple manifestation of natural intelligence."

Overall, I left the talk with plenty to muse on, assuming I will have the time and the brains for this activity.

Reading material:

• “Computing machinery and intelligence” by Alan Turing (1950) (hosted by Hugh Loebner)
• The Turing test entry in the Stanford Encyclopedia of Philosophy.