Artificial Intelligence and “Stupefication”

We’ve touched on two interrelated topics so far: 1) the fact that databases store semantic content via syntactical combinations of data and meta-data, and 2) the fact that formal systems which model intelligent behavior operate under a set of rules that have descriptive but not prescriptive force.

The reason I’ve been addressing these two issues is that I believe stupefication to be a two-component process involving:

1) The distillation of context-dependent human knowledge into a database, and

2) The creation of a complex series of instructions that utilizes the distilled knowledge where possible, and (where not possible) utilizes heuristics — a distillation of context-independent knowledge.

An example of the first case would be Deep Blue’s opening book, which stupefies the game of chess by saying, “If the position looks like that, then do this.” An example of the second case would be an algorithmic implementation of a classic chess adage like, “Control the center,” or “The knight on the rim is grim,” or “Develop your pieces,” or “Seize open files for your rooks.”

But let’s not be deceived by these examples: a computer makes use of these resources very differently from a human being. And the “knowledge” the computer ends up using is of a very different kind than the knowledge we use.

In the following two posts, we’ll deal with each of these issues in turn.