Coming Out of the Close
Artifically Intelligent ComputersThe Closed World should prove interesting to readers interested in philosophy of mind. Edwards1 insights into Alan Turing1s work and the successes and failures of the quest for computers that can be said to be intelligent seems useful, at least to those of us not already expert in that area.
Edwards' brief discussion of neural nets -- computers that presume to operate more like human brains than calculators -- is instructive, but fails to address issues of why we would not want them to run an air defense system. To me, this goes to what should be the main point of The Closed World -- why on earth would we want computers to make moral decisions for us?
In addition, Edwards' description of intelligent computer "agents" in chapter 9, gives a decent if sparse overview of many relevant issues that interested readers could pursue further. What software agents might be used for; ethical problems that crop up annoyingly; what we might have to fear from such software beings; as well as the associated technical difficulties with building intelligent agents. Included is some background, for example, on computers that understand and generate natural human language.