As we enter 2014, and our eighth major generation of gaming consoles, it feels like we as an industry are focused on making everything look and sound more and more realistic. But what about the other components that make gaming the addicting hobby it is? Today I'd like to focus on one of the biggest, and one that seems to be put at the back of the queue constantly with each and every generation of console developers.
While it is true that we have come a long way from the mindless NPC's of early gaming, we're still a long ways off from realistic A.I. I'd say the NPC's in games like the "Elder Scrolls" series, and even "GTA V" are about the limit that we're currently at with AAA titles. Hell even "The Sims" A.I. is pretty advanced at times. Perhaps the most advanced A.I. currently in existence though has to belong to the citizens of "Dwarf Fortress". But what's next? Independent thought? Free will? Skynet? Ok, let's not go that far just yet.
Possibilities, and Limitations
Before we go further, let's think about what we can achieve with our current technology, and what is beyond our scope currently? I'd like to think that we have quite a bit more we can do with the current generation of consoles (PS4/XBONE) regarding A.I. behavior. For instance, one thing I thought of the other day (and actually inspired this whole post) was, what if a game featured NPC's that could give the user a variety of different tasks/quests (depending on a variety of factors), and those tasks would randomly be assigned each time the user plays through the game. I guess you could call it "procedurally generated quest giving". Imagine the kind of crazy replay value inherent in a game designed like that? Each and every player would have their own unique play through experience (I feel sorry for the walkthrough writers though).
We're not quite at this level yet (and I don't see us getting there anytime soon), but what about A.I. that reacted differently to the player depending on their own pre-determined factors? From hunger, mood, temperament, history with the user's character, political/religious affiliations, personal preferences, pet peeves, etc.
For example (and this is a really basic example), imagine an NPC hates the color red. The user's character happens to be wearing red, and the NPC will make a threatening comment, or maybe try to avoid the player (definitely refuse to talk to him, or give him quests). All based on one minor thing that was hardcoded into this particular A.I.
To take it a step further, what if the NPC's all interacted with each other, and hatched schemes, plotted robberies, kidnappings, etc, which the player could be recruited to join.
As you can see this idea can open into an unfathomable wormhole of complications in programming. Which is why I want to ask all of you, what you think about this topic? What do you think would be some fantastic advancements in A.I. behavior this generation (feasible or not)? What would you like to see changed in our current treatment of A.I.? Comment away folks, and thanks for reading this post.