We don't process it poorly, just inefficiently (All things considered)
Video game Computer AI is designed from the top down. That means you break down the system into individual parts and then you design those smallest parts, assemble everything, and you get your overall system. The way humans would build almost any machine, like a TV or toaster or whatever.
Human intelligence on the hand arises from a bottom-up "design". It's the reverse of the way a human would traditionally design a system. Which is why the human brain is so hard for us to figure out, when you look at the thing as a whole it looks like chaos. There is no apparent top-down design that we're used to. Yet somehow there is order and a working machine that allows us to play chess and solve problems on the fly.
We haven't figured out how to "design" true AI, because our first approaches were top-down. We screwed around with that a lot and that's what most computer games use to give you the illusion that you are playing against an intelligent entity. Most computer AI research these days focuses on bottom-up design, using neural nets or whatever. Order out of chaos. If we're ever going to have true AI, this is the only way it will work. A top-down approach will never work, I don't think, except maybe once we have mastered the bottom-up approach first and understand the moving parts a lot better. Bottom-up research is not yet been successful because you have to start simulating really small neural nets. Then larger ones.. and so on. I haven't read about any recent research and my knowledge is probably 15 years out of date, but I would guess we are still really really far away from being able to create working bottom-down engineered AI systems that are anywhere as complex as the brain.