The lights are on
Veteran Member - Level 11
You know what really grinds my gears? Stupid NPCs. Not in a "I is scientistic." kind of way, but in a "Oh look a bush I think I'll walk into that and not recognize it's there." type way.
Close your eyes for a minute. Actually don't because you need your eyes to read this. Imagine yourself playing a game like Rainbow Six Vegas. You find yourself in a nice little nook picking off terrorists with matching cameo pants and AUGs. Soon your spot is compromised, and you have to make a hasty retreat, but you can't. As you are pelted with the few bullets it takes to kill you you think you got stuck on some kind of texture. When you fall you see that your two teammates were boxing you in the whole time, keeping you from escaping. This is just one example of what I'm talking about.
Throughout the history of gaming, innovations are made every day in how stories are written, how characters are fleshed out, how we play games, and how we see them. As much as I welcome these changes and innovations to my fullest, I can't help but think artificial intelligence in games have been sort of trailing. It seems like more often than not the biggest advances in AI are seen in strategy games like XCOM and Total War. Even some of the biggest and most highly acclaimed recent games like Skyrim and Grand Theft Auto V are plagued with unintelligent AI and pathing issues.
So what's the excuse? "It doesn't matter in games like Call of Duty or Battlefield because the challenge is in the multiplayer." What about Pokémon? "It's a game for kids." is usually the default excuse there, but we see a lot of older kids and adults playing it. "They should go competitive if they want a real challenge." That takes a LOT of time and energy. What about fighting games? Same thing. RPGs? "Make the enemies a higher level."
"Man, your Rattata is in the top percentage, Joey? Darn, I'm in trouble now."
I want enemies to actually learn from what you do. Most times baddies have a one track mind that is just "KILL KILL KILL" when you approach them. I want it to go deeper than that. I want them to use the environment to their advantage instead of randomly jumping out and attacking. Say for example, you're walking along a well traveled road. Instead of taking you head on, depending on what you're packing (because having the AI know what they're going up against is a plus), they could attack you from different angles, go for a ranged approach, a sneaky approach, etc.
In the same vein, I want to learn from my enemies. I want to know what makes them tick. Imagine gaming forums still abuzz months after a game's release discussing enemy patterns and attacks, as well as different ways to handle them. That's the kind of smart AI I hope becomes a thing.
What can be done, you ask? Here's an idea: Have a department that is solely focused on building artificial intelligence within a game. Have it be a grueling, gritty, sit around the work table throwing around ideas on a whiteboard type thing. You know, give programming smarter AI as much thought as you would tightening up gameplay or coming up with a good story and characters.
You might be going, "Insane, you don't understand how much time and energy goes into pouring out these AAA games with the best graphics processors can run with the tightest gameplay a controller or keyboard and mouse can give you and stories that are on par with some of the greatest movies and television shows. Why don't you stop nitpicking at the little things and enjoy the bigger product?"
I'll just leave this here.
Well, as games continue to evolve with their consoles and PCs, I can't help but think that AI is going to evolve as well. Maybe not to the extent that I'm hoping, especially now, but maybe in the not so distant future.