Comment by drivingmenuts

How much control over the AI does the end user have? Can I direct the AI in any way I choose (ie., tell it do "evil" of some sort)?