Support NeoGAF

[Tomofdahook17]

I can't be the only one thinking this right?

The fact that A.I. has become so prominent over the past two years but we're still talking about older games that haven't been beaten in terms of NPC A.I. is just very strange to me.

Far Cry 2 is a stand out for me personally because that's the one I've played out of the ones commonly referenced. No other games except Naughty Dog's recent titles such as Uncharted 4 and The Last of Us Part 2 have even come close to impressing me to the degree that this game has.

The other commonly discussed, noteworthy games include (that I know of) Crysis 1-3 (of course) and F.E.A.R.

Edit: Just thought of Alien Isolation as well. An outstanding A.I. system, used for one game and essentially thrown in the trash.

Be sure to mention any other games that you know of that I haven't mentioned.

When do you think the next stand out title will be in terms of NPC intelligence and what developers (if any) do you expect to deliver this quality of A.I. in the future?

 

[March Climber]

 

[Švejk]

The management and testing of outcomes sounds extremely tedious I could imagine.
We'll start seeing trickles of it though. Itsuno stated they're using a form of AI in DD2 for pawn and creature behavior.

Speaking of AI, those Skyrim ChatGPT follower mods are pretty crazy (on YouTube). I would imagine that's basically the future of immersive open world games.

 

[Wooxsvan]

I dont think its quite good enough yet... but could be any day

 

[ungalo]

In my memory Far Cry 2 AI is nothing special, in fact i don't like it. What i remember is that they can shoot you from miles away and they never miss a single shot.

And that shows what's at stake here. AI in video games is not just a matter of programming, it's used as an element of game design. So besides the fact studios probably want to spend resources on other things, there is also the possibility that improving AI will not result in more fun if the entire game is not built around it.

 

[Nankatsu]

Devs reading OP's message:

 

[Guilty_AI]

Because the AI you're speaking of is a completely different thing from the AI that's booming.
Games (that i know of) that used modern AI tech in them:
>Flight Simulator 2020 (synthetic voices and world generation)
>Cyberpunk 2077 (facial animation)
>Spider Man Miles morales (muscle animation)
>Fifa 22 (character movement animation)
>BG3 (Not on the game but on QA tests)

 

[HarryKS]

A.I is a rebranding of stuff that was already out there for many cases.

 

[Danjin44]

 

[MiguelItUp]

I think it's a combination of it's way too soon and maybe even the negativity revolving around the use of it.

 

[daninthemix]

Games take years to make so fads don't just appear immediately in them.

 

[dottme]

Also, there’s probably legal issue. The legal environment is changing a lot around AI and you might not want to risk too much.

 

[Wildebeest]

Ironically, good game AI needs a lot of coordinated human activity to make sure level design, pathing, animation, sound design and other systems work well together.

 

[A.Romero]

You are confusing several types of AI.

What has been booming is GenAI which is fairly recent and would be pretty expensive to implement in a game like say to develop dynamic conversations.

It's very likely that companies are already using GenAI but during the development processes or for stuff like marketing.

The kind of AI that has always been around in one form or another is still there and it's a must for games. Like with any other aspect of game development, companies have different capabilities. Also they have to take into account how AI fits into the game as not everyone is looking to play against super smart bots that will make the average player bite the dust.

 

[Angry_Megalodon]

Some have already given the right answer, they are not the same type of IAs.

Also, by implementing an "advanced IA" they would make games impossible to beat. Or extremely difficult. A Souls without the stairs exploit would give me headaches.

 

[Felessan]

The fact that A.I. has become so prominent over the past two years but we're still talking about older games that haven't been beaten in terms of NPC A.I. is just very strange to me.

Those in the older games are just scripts, not a generative models - so completely different things.
And generative models are hard to incorporate into games as their results are quite unstable and you can't really "fix it" in the way normal programs fixed - so it rule out modern AI as NPC AI for the time being.

 

[Alexios]

You're talking about completely different things, obviously image or frame generation has absolutely nothing to do with NPC or enemy behaviour (nor is just adding chatgpt to NPC speaking so they can discuss about everything somehow good game design).

Machine learning isn't the be all end all solution for such either, it's been used in games since aeons ago with the likes of Counter-Strike's Realbot and Sega's fighting games having modes where the AI learns from the players and what not. Fun AI isn't smart AI.

 

[yurinka]

With this cultural boom of A.I., why is the advancement of A.I. in the gaming industry still so stagnant?

Because games take like half a decade to be made, not 3 months. On top of that, games require predictable and consistent results to be fun and to avoid bugs and must stick to certain age ratings which means some content is censored.

On top of that, the intelligence and skills of the enemies shouldn't be too high because it would mean too high difficulty and would make the games frustrating and not fun. The behavior of enemies is predictable and easy by design, to help players how to act and react instead of being constantly killed by too smart and skillful enemies.

 

[RayHell]

Because Ai was not really Ai since the beginning. It was algorithm for entities to appear intelligent. If we talk about machine learning there's already good papers that can be applicable to the games itself but you're facing other challenges like hardware and specialized workforce. Keep in mind that before it's part of the gameplay, developers will use Ai speed up the workflow of coding, asset generation, upscaling and bugs tracking.

 

[CamHostage]

What gamers think of as "the AI" and what tech industries are employing as "AI" aren't exactly the same thing.

AI in games probably shouldn't be thought of as an artificial intelligence; they're rulesets for how opponent characters should move. (We can get into what modern-day "AI" is in the techworld, but that's complicated to define, so let's just set it aside as separate.) Game characters usually don't really think. They become aware of a condition in a game (the player comes into view, the player has made a noise, the player has attacked a companion, the player has jump-kicked in a forward motion, the player has moved its army to Zerg Rush a fortification,) and then they use specific probability mathematics and pathing algorithms and scripts for what they do in response.

Think of them as chess pieces rather than as opponents. They have a number of pre-defined moves for how they can attack you or defend their space. They are not designed to break the rules. The more moves and the more unpredictable their choice of moves, the more interesting it is to try and defeat them (provided they can be defeated, and that the designers didn't overpower the enemy with moves you don't have countermoves for or tells you can't predict will start attacks.)

For example, in FEAR you hear the enemies calling in on the radio that they're being attacked and see them flanking across the map to get at your position, and you think, "Wow, this AI is so smart!" But, it's not. The game designer is smart, because he thought ahead of time that it'd be cool to write scripts for radio calls and to design subroutines into NPC pathing to have them go around you rather than right at you. The "AI" isn't intelligent, it just has extra moves to use in a fight.

A beginner’s guide to AI: The difference between video game AI and real AI

Welcome to TNWâs beginnerâs guide to AI. This multi-part feature should provide you with a very basic understanding of what AI is, what it can do, and how it works. The guide contains articles on (in order published) neural networks, comput

thenextweb.com

Now, you could use "AI" to train "AIs", but you have to be careful about that. Games are meant to be "games", as in there are designed rules and ways crafted by developers to make it fun. The computer is designed to put up a fight and provide challenging, entertaining gameplay, but it is not designed to win. Players don't like to lose; AIs don't care (and can't care, because they are artificial.) It may be fun to have game characters learn the space and their opponents and adjust with real machine-learned intelligence from player interaction, but that massively complicates the game development process (because how can you test a game that learns from each player, and how can you debug a game when you don't know what it's going to do?) and upsets the balance of map layout and challenge design. If the flying turtles in Mario Bros could learn that you were going to stomp them in order to get over a long ravine, they'd stop flying over that ravine, and poor Mario would be stuck on the other side without a midpoint boost. If everybody in AC could become aware that some nutjob was running around stabbing people in the back and then jumping off roofs into haybales, they'd scream and point you out to the authorities every time they saw straw falling off your tunic. There's an illusion of normalcy that games work with in order to present their highly irregular entertainment experience.

Experiments are being done with "real AI" in games. It can be fun to play a game against a bot trained to do things it learned for itself, or to see a game evolve over time. But take a look at the video below, a simple Tag game where one AI tries to tag and another AI tries to get away and both try to use tools to get their ends met. You can see pretty quickly that the AI may learn things you didn't expect and, as a player, may not be able to contend with. AI can be fun or it can be frustrating, and game designers are the craftsmen of how fun/frustrating gameplay is; put that in a machine's hands and you can't be sure what you'll get...

 

[ClanOfNone]

One of the old Unreal Tournament games had pretty good bot ai, at the right settings they felt very “human.”
On the highest settings they were ruthless aimbot terminators.

 

[Fbh]

Because it feels like except for a few devs, most of the AAA industry these past generations has been focusing on muh graphics.

 

[Rudius]

Because the people making games are not the same as used to be.

 

[StueyDuck]

going completely on nothing than what i just personally see in the industry as an outsider (i have no evidence to back this up, this is just conjecture). It's probably due to most of gaming and it's resources going to the art department and the community managers side.

something tells me there are far less developers working on games than those who make the games look pretty or write the script. I'd take a game that's simplistic in visuals but pushes boundaries in engine tech anyday.

 

[_Ex_]

Improving enemy AI would only raise the difficulty, and most gamers don't like difficult games. Developers would rather use AI to produce less expensive assets for games instead.

 

[Kssio_Aug]

Improving enemy AI would only raise the difficulty, and most gamers don't like difficult games. Developers would rather use AI to produce less expensive assets for games instead.

I don't really buy that. I would rather play a game where the enemies were really smart and constantly surprising me, but that could take about 10 hits to kill my character, than a game with a dumb AI that repeat the same pattern over and over and is able to kill with 2 hits.

It's all about balancing. With a smartest AI, you can give the player the edge with different tools.

I think the real issue here is that they know that graphics sells a lot more. It's better to invest on making a good looking game that will standout in trailers and such, and everyone in gaming communities will be talking about how amazing it might be, than investing their time and money in improving AI while people would be criticizing it's graphics saying it doesn't look like a next-gen game.

 

[SHA]

Risks/mind shift, it's necessary.

 

[buenoblue]

Because AI is just a buzzword with nothing artificial or intelligent about it in any way

 

[FunkMiller]

The ‘AI’ that lets you make a picture of Taylor Swift dressed as a sexy waitress is a very different thing from the ‘AI’ that creates NPCs that function in believable and humanistic ways.

One is easy. The other is really fucking difficult.

 

[sendit]

Because AI is just a buzzword with nothing artificial or intelligent about it in any way

not true. generative AI is both artificial and intelligent.

 

[Tomofdahook17]

What gamers think of as "the AI" and what tech industries are employing as "AI" aren't exactly the same thing.

AI in games probably shouldn't be thought of as an artificial intelligence; they're rulesets for how opponent characters should move. (We can get into what modern-day "AI" is in the techworld, but that's complicated to define, so let's just set it aside as separate.) Game characters usually don't really think. They become aware of a condition in a game (the player comes into view, the player has made a noise, the player has attacked a companion, the player has jump-kicked in a forward motion, the player has moved its army to Zerg Rush a fortification,) and then they use specific probability mathematics and pathing algorithms and scripts for what they do in response.

Think of them as chess pieces rather than as opponents. They have a number of pre-defined moves for how they can attack you or defend their space. They are not designed to break the rules. The more moves and the more unpredictable their choice of moves, the more interesting it is to try and defeat them (provided they can be defeated, and that the designers didn't overpower the enemy with moves you don't have countermoves for or tells you can't predict will start attacks.)

For example, in FEAR you hear the enemies calling in on the radio that they're being attacked and see them flanking across the map to get at your position, and you think, "Wow, this AI is so smart!" But, it's not. The game designer is smart, because he thought ahead of time that it'd be cool to write scripts for radio calls and to design subroutines into NPC pathing to have them go around you rather than right at you. The "AI" isn't intelligent, it just has extra moves to use in a fight.

A beginner’s guide to AI: The difference between video game AI and real AI

Welcome to TNWâs beginnerâs guide to AI. This multi-part feature should provide you with a very basic understanding of what AI is, what it can do, and how it works. The guide contains articles on (in order published) neural networks, comput

thenextweb.com

Now, you could use "AI" to train "AIs", but you have to be careful about that. Games are meant to be "games", as in there are designed rules and ways crafted by developers to make it fun. The computer is designed to put up a fight and provide challenging, entertaining gameplay, but it is not designed to win. Players don't like to lose; AIs don't care (and can't care, because they are artificial.) It may be fun to have game characters learn the space and their opponents and adjust with real machine-learned intelligence from player interaction, but that massively complicates the game development process (because how can you test a game that learns from each player, and how can you debug a game when you don't know what it's going to do?) and upsets the balance of map layout and challenge design. If the flying turtles in Mario Bros could learn that you were going to stomp them in order to get over a long ravine, they'd stop flying over that ravine, and poor Mario would be stuck on the other side without a midpoint boost. If everybody in AC could become aware that some nutjob was running around stabbing people in the back and then jumping off roofs into haybales, they'd scream and point you out to the authorities every time they saw straw falling off your tunic. There's an illusion of normalcy that games work with in order to present their highly irregular entertainment experience.

Experiments are being done with "real AI" in games. It can be fun to play a game against a bot trained to do things it learned for itself, or to see a game evolve over time. But take a look at the video below, a simple Tag game where one AI tries to tag and another AI tries to get away and both try to use tools to get their ends met. You can see pretty quickly that the AI may learn things you didn't expect and, as a player, may not be able to contend with. AI can be fun or it can be frustrating, and game designers are the craftsmen of how fun/frustrating gameplay is; put that in a machine's hands and you can't be sure what you'll get...

Thanks for this detailed explanation!

That video is also very fascinating, I think if anything, that's the type of NPC capabilities you would want in a multiplayer/co-op game with bots that could be played for hours and hours. That way you would never see the same behavior twice, and if you did you would know that the NPC's are still working towards better solutions and outcomes.

 

[ChuckeRearmed]

I think it is two things

• People are stupid and tired and don't want to fight again smart AI (too hard, too boring, too smart AI etc.)
• There is less competency in gamedev as almost every is the assets created by other people