I think, even in principles there are some fundamental issues. The first is that our views on various political issues change over time. An algorithm to duplicate us would have to be updated constantly to use past decisions to predict futures ones. This may not be possible to do accurately, even in theory.Ulrich wrote:Yes. Your point and the Red Dwarf excerpt which illustrates it show some drawbacks. But I think it's interesting to forget about current technology and concentrate on the princple being discussed. i.e. In principle, could it ever be a good idea to try, in some way, to encapsulate our wants and needs, for the purpose of voting, using stored data and an algorithm of some kind linking that data?
The second is that the various political issues change over time. Voting republican carries a very different implications from what it did a few years ago. Voting in the UK for almost any poilitical party now carries a very different set of priorities from the last election. It's not clear to me that these differences can be captured in an algorithm, since the information of which party supports which values or positions isn't known, even by the parties themselves.
What I would say is that if we are going to go with Algo voting, there is no real reason to involve the voters at all. Just calculate their views centrally, from sampled data, and do away with voting altogether. It's a functionally equivalent proposition, and much cheaper.