to build a decent artificial inteligence (and i mean decent as not to make stupid mistakes not that it will be a true challange) it would probably require at least 200k lines of proper coding plus testing bug hunting and so on so at least a team of 20 people over 1 years work wich could be lost by a mistake or then the code could be stolen near completion (yes there are people in this world who will do almost anything for profit)

so for a good artificial inteligence you would require 10 times more work for it and for a trully amazing artificial inteligence (one that "plan ahead" deal with the economy, army management, public construction, diplomacy, population control/managements, tactical movement strategic thinking and batle fighting operations )

thats 2 much of a daunting task imho, most games go with fresh aproaches/concepts cause going with a true ai wich isn´t visible only felt after a bit of gaming doesn´t sell and it would take 2 much work for litle profit revenue. as i said ca could do this to prepare for a future game competition when a worthy opponent finally arrives and quality and every litle detail will matter.

just my view on this subject and only for the artificial inteligence i don´t even go into skining or other stuff.