I saw Brian Eno talk years ago and he said he thinks we've used computers for the wrong things. He argued that instead of using computers to plow through piles of data to arrive at uninteresting results we should emphasize the ability of computers to generate surprising results from very small seeds of data. Examples of this would be fractals, algorithmic music etc.

There are a fair number of people working on algorithmic & procedural art but a lot of it is pretty dry. If you can find a way to use a computer to create something that really moves people you can probably find a market.