Benefits of higher performance Console CPUs (Game AI) *SPAWN*

That's a problem with AI in general, such as perfect aiming with perfect deflection on moving targets.
This also becomes more problematic for networked gaming as it's harder to sync such dynamic simulations over a network.

Although my preference for next gen would be the end to invisible barriers and for all terrain interaction to be physics based, so every clearly achievable movement, every climb and jump that the play knows the character could make, is possible. Nothing worse than trying to escape up a rocky outcrop into cover only for there to be a no-go barrier! It became laughable in Uncharted where, on the one hand Nate can make superhuman leaps to reach impossible areas, and yet on the other he couldn't even climb over a rock or up a small hill because he wasn't supposed to.
Indeed, though, I can only imagine this future would end up with game worlds placed inside large craters or valleys to ensure players not getting outside the play area instead lol...
 
Indeed, though, I can only imagine this future would end up with game worlds placed inside large craters or valleys to ensure players not getting outside the play area instead lol...

Yes, this is more of a level design issue than anything else.

I think AI in a few games is already pretty good, the enemies in Halo, Killzone or Uncharted have a range of interesting behaviours and are challenging to fight.

Animation is probably a bigger issue than AI to get right if you want to have realistic characters.
 
Last edited by a moderator:
Agree with more difficult animation for realistic characters. I noticed the animation flaws in the KZ4 stage demo almost immediately. I miss KZ2. ;-)
 
Agree with more difficult animation for realistic characters. I noticed the animation flaws in the KZ4 stage demo almost immediately. I miss KZ2. ;-)

Yeah, i'm playing KZ3 at the moment and I can't count how many timesI see something like Rico try to climb down a ladder where he starts by grabbing some invisible rungs before climbing down to the ladder itself,
 
polygon milo and kate
http://www.polygon.com/features/2013/3/21/4063508/milo-and-kate

" It used smoke and mirrors to cover its weaknesses, but a player could stand in front of Milo and ask questions and tell jokes, and Milo could recognize them and respond as if he understood. He didn't always, however, which lent the software an occasionally awkward tone."

at lease from the "smoke and mirrors" demo, the AI already allow a conversation that not awkward like skyrim
 
Last edited by a moderator:
At first, it reminded me of the PS3 launch "advertisement", Eyedentify. Then it reminded me of the autonomous catbird in The Last Guardian. The track record seems pretty bad for an autonomous AI game. The simple ones like Tamagochii and Nintendogs sold very well, so the AI character may not need to be super smart or voice controlled for entertainment.

Today, people use Siri, Google Now and perhaps Kinect to tap on the Internet's vast knowledge. In these cases, we don't have to render the AI character. Just make the voice recognition less frustrating and more consistent for a solid start.
 
I'm getting a sense of déjà vu. I'm sure this discussion has been done before.
 
The AI in any game has only one purpose: make you feel good when you win.
That's basically it, therefore "game AI" is a matter of psychology and sociology and not of machine learning or any other "real AI" stuff.

There's also the ai for npc's and other game related things that have no bearing on whether or not you win or lose, but are there to make the game look and feel like a real living world, as well as to make the world feel like an interesting place to be. Right now ai is woefully inadequate for that. Worse yet, these sorts of things are for the most part still being handled the same way today as they have been for the past 15+ years. There have been little to no advances on ai in that realm which is why games today still feel scripted and fake for the most part when it comes to npc's and other ancillary game components.
 
Oh I completely agree. There's better games in that regard, but most ancillary AI is more than just underdeveloped. In GTA, they're just there to give the city people to interact with. And by interact, I mean shoot, drive over, or ignore. I always found it funny when people name Skyrim is having better AI, since it has a schedule. It's just as bad, imho, because the interaction is no different. I rather think it is, as Joker said, even more scripted. Come in the evening, the smith goes from his store to his home... wait for the morning and he's back. That's it. And as a gamer, I find that annoying more than anything, because I need to wait for him to return if I want to buy something.

But... I actually can't say what needs to change, tbh. Making them do more stuff will lead to more annoying situations where you can't find the NPC you're looking for. But on the other hand, you could couple this with gameplay. Again Skyrim. The smith goes out to find more wood or make deliverances. He posts a note to his store saying he's out and where. Your UI gets a marker telling you where he went. And if he's out in the woods, he might have been attacked by whatever and you can save him. Still just as scripted as before, but with added gameplay attached to it. Not sure if it works or not.
 
There's also the ai for npc's and other game related things that have no bearing on whether or not you win or lose, but are there to make the game look and feel like a real living world, as well as to make the world feel like an interesting place to be.
That's true, but it's also a very niche set of games that'll benefit from this. There's good reason why the state of the art in AI hasn't improved much in 15 years - it's very complicated and processor/RAM intensive and fairly limited in application.
 
That's true, but it's also a very niche set of games that'll benefit from this.

Please elaborate on which game types you think fit in this niche category. There are NPC's in the majority of games I play. Not all of the same type, but all need AI routines.
There's good reason why the state of the art in AI hasn't improved much in 15 years - it's very complicated and processor/RAM intensive and fairly limited in application.
Is AI RAM/processor usage proportional to advances in hardware like, for example, rendering is?
The number of concurrent AI in any given scene might be trending upward (I don't know, haven't really looked into it) so I can see higher resource usage in that scenario, but as more system resources become available with every generation surely the percentage allocation to AI is shrinking?
Can you give any concrete examples of resource usage for AI?
 
Please elaborate on which game types you think fit in this niche category. There are NPC's in the majority of games I play. Not all of the same type, but all need AI routines.
Living world games require non-combat AI as Joker describes. For most games, AI is for opponents, from shooters to racers to sports games to platformers to beat-'em-ups to dungeon crawlers, etc. The vast majority of games only need AI enough to present an interesting challenge, and in most of these games, online provides a 'cheap' solution to a challenging AI. I agree that offline bots etc. can do with being smarter for those that play offline, but I can also see why devs won't be investing heavily into AI when they have an easier solution at their disposal.

Is AI RAM/processor usage proportional to advances in hardware like, for example, rendering is?
The number of concurrent AI in any given scene might be trending upward (I don't know, haven't really looked into it) so I can see higher resource usage in that scenario, but as more system resources become available with every generation surely the percentage allocation to AI is shrinking?
Can you give any concrete examples of resource usage for AI?
I can't give any specific data. I can say that you end up with exponential and messy growth akin to raytracing or physics modelling. For each agent you need to test scenery for possible interactions and other agents, and quite often ray/path calculations. eg. To fire a gun with leading you need some non-trivial maths. To predict where they might change direction to also requires maths. 100 agents all testing rays against each other and scenery gets inefficient. Also the basic premise of decision making boils down to a load of conditionals, meaning lots of branches and random access.

Let's look at something like an Elder Scrolls game. You have an agent that can move objects, and want this agent to set the table for dinner. They need to evaluate all the local objects to decide if it should be on the table or not, test every object being placed to make sure there's room (if you put a jug in the way of where they are going to put the plate, they need to stop their action and respond), respond to changes in the room state to see if anything's making a mess of their task (you putting knives and forks back) and to determine when the task is complete. They need to respond to a knife not being present because another AI agent had taken it out before to use for some job, and needs a start a search that doesn't lead them to the store to buy a new knife or wandering the woods performing a simple locality test for the presence of a knife. And with the interactions becoming more complex, the chance for bugs and errors increases exponentially. There's a good chance IMO that you'll hit artificial madness before you hit artificial intelligence. We've seen that already with NPCs walking on the spot into fences etc. ;)

Considering that most games don't need anything like this level of complexity, just needing NPCs to cover, shoot, not stand still getting shot at, and have some sort of tactics (or other behaviours necessary for the game), it's not surprising that there's no significant investment in AI and it's down to the open-world games to create their own.

And here's an example of poor AI bringing entertainment value! http://www.express.co.uk/entertainm...r-GTA-5-PC-Mount-Chiliad-GTA-5-Taxi-challenge

I think GTA is a great example. Improving the visuals is a far, far simpler task than creating intelligent virtual people fully aware of their surroundings and responding with simulated emotions and convincing interactions to each other.
 
That's true, but it's also a very niche set of games that'll benefit from this. There's good reason why the state of the art in AI hasn't improved much in 15 years - it's very complicated and processor/RAM intensive and fairly limited in application.

I dunno if I'd call them niche, I think often the top selling games in a given year are open world games like Watch Dogs, Assassins Creed, Skyrim, GTA, etc and then the rest are typical shooters. But I'll leave it to y'all to decide if they are niche or not. I would add that open world games tend to consume more of a gamers hours than other types of games do though, which makes ai all the more important there especially when you hit the 50+ hour mark on a game. Right now even on a high budget game like GTA the ai you get is like this: http://imgur.com/gallery/vTgLH . I'd agree that it's a complicated task to solve, but I was countering the point that the only purpose of ai is to let the player win and feel happy about it.
 
I think GTA is a great example. Improving the visuals is a far, far simpler task than creating intelligent virtual people fully aware of their surroundings and responding with simulated emotions and convincing interactions to each other.

This is what WATCH_DOGS promised to deliver and while the AI can, in some ways, be more sophisticated and varied in terms of how pedestrians interact and react to each other, it falls far short of GTA in other ways. But even in a modern GTA game you tend to feel the centre of attention and the only instigator of chaos. I'd like to see peds getting into disagreements and fighting, or pets carjacking other peds. I remember watching into a coffee shop in WATCH_DOGS and some guy was hold it up. I remember the gang wars of San Andrea and Vice City Stories.

This is barely even AI at this point, it's just more varied events. It's not like a particular ped in a city of 100,000 peds is undergoing artificial anguish about losing his artificial job so is unable to pay his artificial mortgage for his artificial house so robs a store.

edit: typo - peds not pets :nope:
 
Last edited by a moderator:
I dunno if I'd call them niche...
Counting them as proportion of total games made and so number of devs interested in advancing the art. I agree with your point that hours played by gamers is a better metric of whether they're niche or not for gamers, although it'll still be a fairly small percentage.

Or putting it another way, if you set out to create the greatest AI engine of all time and license it to living world games, you'd be selling to a handful of developers and need to charge a gazillion pounds per license. ;)
 
This is barely even AI at this point, it's just more varied events. It's not like a particular ped in a city of 100,000 peds is undergoing artificial anguish about losing his artificial job so is unable to pay his artificial mortgage for his artificial house so robs a store.
Modelling that sort of psyche in an agent isn't that hard. The problem is how they go about acting on a decision. Basically at the moment we can only choose from a number of prebaked options. For true AI, you'd need completely open world interaction with objects having properties and derivable actions that can be strung together into actions. Well we haven't even got that for the player yet, let alone AIs!
 
More like make you feel good period. If a game is challenging, keeping things fun and non-frustrating even when you're being wrecked is perhaps the most important goal of all.
The thing that disappoints me the most about AIs is that in most games they are just scraps of meat for the lions, they move around mindlessly. I think games should adapt the AI when increasing the difficulty levels, rather than add more hit points and attack to them.

At least Halo and GeoW baddies try either to hide or make live difficult for you, but that's not always the case.

I remember watching an Assassin's Creed video (AC IV iirc), where the AI is clearly meant to make the gamer feel like god, and a guy had killed all the enemies around, yet the last one standing was there patiently waiting for the player to kill him instead of running away.

He showed how ridiculous the AI was, he beat all the AI guys -like 10 people surrounding him- without moving the character at all and pressing just a button from time to time... :(
 
make the game look and feel like a real living world

In most games it's not the goal. There are niche simulation games that are like that, but majority of games strive to be fun and entertaining and not "close to reality".
Even movies, which can render each frame for hour or so, are more about "fun to watch" than "close to reality".
I would argue that improving game AI using simulation algorithms is a dead end, it makes games complex and does not add to fun value.
Heck, how you can even talk about "real world" humans if the physics in games is as far from "real world" as it gets. Properly solving "a lot of rolling barrels" would require a lot of computing resources, much more than any current PC has right now.
 
Back
Top