In that specific context - of generating idle chit chat, sure. But is it ever going to be capable of generating the crucifixion quest from CP77, or Guild quests from Skyrim or the Festers Blue Star Bottlecaps from FONV?
or is it going to be more A New Settlement Needs Your Help from FO4, or Dunk the Shape / Kill X Enemy Ys from Destiny 2? which, yknow, we already have.
Generating idle text does not a great game make. Especially when you could just write it better.
And that’s not to mention the impact on the VO actor - who is unlikely to want to sell the IP to their voice
I am actually working on something for the quest generation problem. It is still in the experimental phases, so who knows if it will bear fruit, but don’t sell the concept short.
I remain politely skeptical. I’m not the least technical person- but also not a dev - but this AI has to create multiple NPCs that say sensical things, in a narrative form, in a reachable location, in a playable architecture and geography, using themed assets, realistic and not over-/under- powered rewards… draw, plot and arrange said assets, actors, cues, generate speech-to-text and assign the correct asset to the correct cue/trigger — all of which seem to me to be beyond the reach of AI/ML models at the current point in time, or else subject to multi-hour loading and generation times.
Then there’s the issue of if you’re generating assets for the engine, and it needs a filesystem to store those assets, is it not incredibly easy to create massive security holes? An attacker looks at the program, see it generates and FBX or OBJ and can use that as a security hole to inject malicious code.
Also, doesn’t engines like Unity, Godot, need to compile these assets and process them? It’s beyond my technical knowledge but you can’t edit game assets on the fly, right? Like I can’t just open up MYGUN.TEXTURE and paint it blue and now I have a blue gun without closing the game, right? How do you work around that?
You can change assets on the fly, yeah. Usually with stuff like making a gun blue you’d just load another texture and apply it to the material. It really depends on what the game is designed to do.
For example a game where all the lighting is baked would have issues if certain parts of the level were changed in real time because you’d need to rebake the lighting (or add some dynamic lights specifically for certain objects)
Stuff like creating a quest in real time to the extent of hand crafted quests doesn’t sound like it’s quite there yet but there doesn’t seem to be a technical limitation there other than what AI can do and how to refine it to do that in an interesting way.
You never know but it still feels a bit early considering how little has been done so far.
Not to belabor the point here, but in a discussion of “can AI do this” [now/soon] - saying “if AI could do this… then it could do this - but it cant - but it might” doesn’t seem to really counter my point that the next 5 years being full of empty promises about the potential (but not actualization) of AI.
On your first point, no, an individual AI is not, and never will be, capable of doing all of those things. What is will be is an analog to how the human brain works. You don’t see, hear, move, and process the words of a conversation you have while walking down the street with a friend using the same pieces of the brain. The occipital lobe, auditory and locomotive sections of the somatosensory cortex, and language center of the prefrontal lobe handle the parts independently of each other, then the information is brought together and presented to your conscious mind. An AI-driven questing system would have multiple specialized AIs that worked together to generate it. So a model which analyzes the current state of the player to determine valid reward thresholds and quest objective difficulties, another one which maps the current world lore to make sure that the quest fits into the world state, another which fills in all of the dialog based on NPC background variables, then a final AI which is trained to look at the outputs of the others to resolve conflicts. Finally, an AI voice synthesis can round out the experience for players. All of those can run in parallel and can use quite a few metrics from player interaction as feedback for refining the training.
To your second point, most of the aspects of a quest are rather small and can be stored in memory. The rewards can get interesting. If they are a world object, procedural modeling can go a long way to making it so asset generation is not necessary. If it is perks or traits of some variety, this could be something generational which uses keyword detailing to create the parameters for the trait. Generation and storage of details and items are not really much of an issue.
As for the engine questions, all of them can process geometry, textures, and text from memory or new files on disk. If something needs to be compiled, then it can be compiled on the fly. Again, individual assets are pretty lightweight and would not require a lot of processessing.
Another speed-up would be to pregenerate details of the quests rather than attempting to do it all using a just-in-time implementation. The game could generate the parameters for the world for NPC’s in town when you load in, starting with the ones closest to the player position and progressively iterating over them in the direction of travel. All you need to do is have details ready for the “chat bot” portion of the interaction by the time the player is capable of reaching any given NPC. These are the boundaries of what is possible so not as heavy as generating the whole thing. Then the rest can be filled in while the player talks with them.
The biggest issue I see is continuity error hardening. Making sure that all of the NPC’s worlds are consistent with each other and nobody makes changes that break the world for other NPC’s. That is specifically what I am trying to work on.
I’ve been hearing promises of the human brain being replicated by a PC since the 1980s, so again, politely, I consider that hyperbole/marketing gumpf until we see a working model.
I don’t particularly care if “AI” means a single model, multiple models, or multiple models banded together to appear as a single model or vice versa.
I didn’t realize you can just chop and change models on the fly, but taking those and similar issues as read (or at least probably solvable with modern day tech) — that leaves us with your multiple-AIs with specific functions.
Now I’m not saying it’s theoretically impossible, but i am asking: will you have a working prototype that can be run on a consumer home pc in the next 5 years? Or, are you, as I am very keenly aware of, simply doing what I stated in my first post: being a tech start up promising eventual, incremental process as product features?
Because my experience is, to generate a flat image from a model takes at least 20 seconds, not to mention 3D models with collision, mesh, animations etc. And 20 seconds is considered a long loading screen by modern standards. Gamers expect entire cities/planets to load in their game in under a minute.
Are we even close to generating even an untextured room with a single untextured T-pose NPC with no cues, triggers or animations from AI? Or, would it simply be using a language model to obfuscate the current and standard process for generating those assets to the user, when actually it’s just loading them from RNG.
Are we close to generating an untextured room with a single untextured T-pose NPC?
This is unnecessary. The environment and characters can all still be built and rigged using modern hand modeling and set design techniques. The get here would be the animations. All animations, and I mean all, are just a bunch of splines and interpolations in the engine. Curve fitting is something that AI’s have been getting trained to do for 50 years. It is a solved problem. What is not solved, and the reason we don’t see a ton of AI animation tools already is a lack of sufficiently large sample sets of good animations attached to a variety of rigs and a method of training the AI to be environmentally aware. I imagine the latter could be solved using computer vision techniques with virtual sensors like what is used for Crowds. It is actually a hypothetical problem that I am planning on tackling after my current project.
Or would it be using a language model to obfuscate using an RNG?
This is almost insulting to the entire field of AI development, not to mention the mathematical fields of Probability Theory and Statistics. While AI models do function on Game Theory and Probability Theory models and are probabilistic, they are most definitely not any form of RNG. If you have that mindset, I suggest you do some more reading on the topics in scholarly publications or textbooks. Skip the pop-sci articles and blog posts.
As to your questions on timing, it may be this year, it may be 20 years. 2 years ago the idea of making photoreal images with AI was a pipe dream, then along comes Stable Diffusion and in less than 18 months we have gone from making passable images of a cat to an almost fully art-directable toolset capable of creating coherent videos. This is an amazing progressive leap and AIs in general have become more diverse because of it. It is a testament to just how fast something like this can grow given the right FOSS architecture and public interest. My guess is that it is closer to the 2 to 3-year range for a coherent world-building AI. That is not modeling, rigging, or animating, just the textual stuff. Story, relationships, lore, history, etc. My first tool I am trying to build is a world-building assistant for TTRPG GMs if that gives you a clearer picture of what I’m talking about.
Finally, the geometry generation times. You mention your experience, but I am struggling to pin down exactly what that entails for you. I mentioned using procedural modeling and having an AI that decides the parameters of the procedural when doing the modeling. In this setup the modeling is all mathematics and is done instantly. It can even include procedurally defined animations and affects that are able to be generated on the fly in microseconds. If you have 3D experience, I would suggest checking out Houdini ( https://www.sidefx.com ) to get a better grip on what is possible with proceduralism. They also have tools for doing rigging and animation that an AI could directly interface with, which can be utilized in game development realms as well as VFX.
You misread me when i said RNG - I was saying using a language model to appear as if it’s generating a quest on the fly, but actually its just picking Kill [5-30] [Goblins / Demons / Beasts / Mechs] from a list is not “AI generated quests”
And you misunderstood what I was saying about being a Dev - I said I was not a dev.
I know the following style of reply is a bit rude - but I want to highlight how I said that much of what tech companies are promising with AI is an empty promise that will only lead to semantic satiation of the term “AI” leading it to become a term of derision
can be built with hand modeling techniques
so not AI.
lack of sample sets
so AI can’t do it (yet)
[no] method of training the AI to be env aware
so AI can’t do it (possibly ever)
it may be 20 years
so unlikely to be a reliable marketing promise to consumers in this generation or the next
my guess is it’s closer to 3 years
sounds like marketing bumpf
…just the textual stuff
so not AI generated quests in game as originally promised…
Look, I really don’t mean to pick a fight with you - but saying “I’m working on AI generated quests!” to mean “I’m working on using an existing LLM to create text-based lore entries on my world” is a very different expectation to reality ratio.
Your response style is not rude. It is disingenuous confirmation hunting cherry picking. I addressed everything in your response in detail, but you latched on to the qualifiers and clarifications because those are the parts that satisfy your confirmation bias, even though in the full context they mean quite the opposite. I also am recognizing that you are not a dev, not a game designer, not an AI architect nor mathematician, or a computer scientist. All of this means that the scope and breadth of your understanding of the topic you are attempting to belittle and demean is myopic at best. It is obvious that you do not understand what a “quest” in a game is nor what it takes to craft or write one. It is clear that you have 0 understanding of how LLM or AIs as a larger topic function or generate information. You deign to belittle the work that I am doing on this topic without asking a single question or clarifying a single detail from me. I’m sure you are getting a nosebleed for your perch on Mt. Dunning-Kruger. I can see you from my position on the adjacent slope. I will humor you though, since you seem to at least care about the topic, even though you seem utterly incapable of recognizing when something should be informative and educational.
What is a “quest” in a game:
The abstract of this article covers some key concepts of both current anf potential future paradigms of quest design in video games. Currently, quests, even handcrafted ones, consist of a list of tasks. Lists of tasks are text. Well-crafted ones include lore hooks, personal interactions with a variety of NPCs, tasks that are meaningful to the world, and rewards/results that either expand the player’s capabilities or draw the player into the world and the story. Components needed to craft a quest are a means of initiating the quest for the player (quest givers, discovered information like a letter or journal, an event that happens in the player’s presence), dialog, goals/tasks, and a location. For a well-made persistent world, locations, and NPCs are persistent as well. They don’t need to be AI-crafted on the fly. They should have been created when the game was made, or at least some period long before the players interacted with them. What can be crafted as a just-in-time or an emergent experience is the interaction with the player. Dialog with lore hooks that give the quest grounded footing in the world. Interactions with other NPCs which are based with the relationships of the NPCs with each other instead of just the player, etc. All of this is textual in nature. Dialog, relationships, interaction. The secondary stuff, like animations, could be handled in a lot of clever ways. They could be hand-crafted by animators and an AI could be trained on what animation links to what interaction type. The AI then selects and blends the canned animations when it is interacting with the player. And before you have something to say about that, it is literally how animation is done now by people. I absolutely have authority on this topic as I have a graduate degree in animation and visual effects.
The AI I am working to create currently is being designed around the goal of making a consistent world with complex relationships and personalities so quests can be written that are self-consistent and coherent with the world. It is not worried about writing dialog or generating rewards, just making a world and quests that are cogent and have minimal plot holes.
What isn’t part of a quest?
Geometry, models, textures, particles, VFX, rooms, people, trees, animals… basically anything that has physical form or defines a physical characteristic of an object. These things may be involved in the completion of the tasks themselves, but they are all separate from the quest. As an analog, your wife hands you a grocery list and asks you to go to the store to pick up the things she needs for dinner. This would constitute your bog-standard fetch quest in a video game. The quest is to obtain the listed items. All of the items exist independently of the quest. The grocer and cashier at the till do as well. As does your wife. The only thing that the quest consists of is the request which is made, and the list which is requested. Everything else does not need to be created, it already exists. I hope that clears up your misunderstanding of the topic.
You can erase the entire concept of modeling or texturing from your mind in relation to an AI creating quests. I mentioned it because I recognise that some objects may need to be created. Perhaps your wife asks for a product which the store did not contain already. So in a gaming context, this should be cleaned up so the quest is able to be completed. The product needs to be created. You can place limits on the list of items so they are all canned goods. Now you have a rudimentary prototype object that can be anything. Canned beans, sure. Canned orc tongue, why not? All that is needed is a label and a spot on a shelf. SD can create a label in a few seconds while the player moves from his home to the store and an observation-aware procedural system can stock it on the shelves when no players are looking and inform the AI driving the cashier of the existence and location of the item. Nothing needed modeled. Hell, all that was really needed was a procedural text generator to make a label for the can, no AI was needed for that. What the AI is needed for to make the trip to the grocery feel right is the request from your wife, the conversation with the cashier at checkout, and the smile and thanks from your wife when you got home with the items.
If more modeling is needed, procedural modeling takes care of that. Artists create prototype objects with fixed bounds on parameters and an AI is trained on how to set those parameters, so whenever a quest needs an item, the quest generation AI requests the type if item and the context in which it will be used, and the modeling AI interprets the request, prepares the correct prototype, determines the contextually appropriate parameters for it, then places it where the quest giver needs it. No muss, no fuss. Eventually, there will be SD analogs for 3D geometry and generative modeling. It is a field of active research at IBM, nVidia, AMD, Meta, and many others. It will happen, and it will happen sooner than anyone will be ready for it, much like SD. But for now, it is overkill. AIs can use procedural modeling to adequately furnish and populate a world without much overhead. I think I may play with that as well. Rip the character creator from the likes of Skyrim or Starfield and train an AI on faces and the character creator parameterization and have it go wild. The AI is still creating the characters, it is just using the same tool as the player to do so.
There are already mods that add voices to mod-scripted lines for Skyrim and Fallout 4. As well as a joke mod where the author took all of the recorded voice lines for Deep Rock Galactic(DRG), ran them through AI translators 40 times, then had an AI record the end result using the intonation and inflection of the characters in DRG.
However, the quotes from MS in the article provide an insight into their plans.
Replace all localization teams.
Replace all QA teams with AI that just run the level infinitely.
Interestingly, both of those show a fundamental lack of understanding of what a LLM can do. Yes it can do basic translation, but it fails on context in translations. E.g:
The French translation “you are a fool” can be “tu es un imbecile” but can also be “vous etes un imbecile” depending on the relationship context of the people talking.
And asking AI to replace QA entirely… oof, I guess I know to avoid MS games at launch from now on. Will be a lot of bugs.
I doubt it. AI is actually useful for games. I’d love a Skyrim where there were infinite unique npcs who don’t repeat dialog on a loop.
In that specific context - of generating idle chit chat, sure. But is it ever going to be capable of generating the crucifixion quest from CP77, or Guild quests from Skyrim or the Festers Blue Star Bottlecaps from FONV?
or is it going to be more A New Settlement Needs Your Help from FO4, or Dunk the Shape / Kill X Enemy Ys from Destiny 2? which, yknow, we already have.
Generating idle text does not a great game make. Especially when you could just write it better.
And that’s not to mention the impact on the VO actor - who is unlikely to want to sell the IP to their voice
Will it ever be capable of that? Most certainly yes.
But we won’t ever get there if nobody does the first step.
I am actually working on something for the quest generation problem. It is still in the experimental phases, so who knows if it will bear fruit, but don’t sell the concept short.
I remain politely skeptical. I’m not the least technical person- but also not a dev - but this AI has to create multiple NPCs that say sensical things, in a narrative form, in a reachable location, in a playable architecture and geography, using themed assets, realistic and not over-/under- powered rewards… draw, plot and arrange said assets, actors, cues, generate speech-to-text and assign the correct asset to the correct cue/trigger — all of which seem to me to be beyond the reach of AI/ML models at the current point in time, or else subject to multi-hour loading and generation times.
Then there’s the issue of if you’re generating assets for the engine, and it needs a filesystem to store those assets, is it not incredibly easy to create massive security holes? An attacker looks at the program, see it generates and FBX or OBJ and can use that as a security hole to inject malicious code.
Also, doesn’t engines like Unity, Godot, need to compile these assets and process them? It’s beyond my technical knowledge but you can’t edit game assets on the fly, right? Like I can’t just open up MYGUN.TEXTURE and paint it blue and now I have a blue gun without closing the game, right? How do you work around that?
You can change assets on the fly, yeah. Usually with stuff like making a gun blue you’d just load another texture and apply it to the material. It really depends on what the game is designed to do. For example a game where all the lighting is baked would have issues if certain parts of the level were changed in real time because you’d need to rebake the lighting (or add some dynamic lights specifically for certain objects)
Stuff like creating a quest in real time to the extent of hand crafted quests doesn’t sound like it’s quite there yet but there doesn’t seem to be a technical limitation there other than what AI can do and how to refine it to do that in an interesting way. You never know but it still feels a bit early considering how little has been done so far.
Not to belabor the point here, but in a discussion of “can AI do this” [now/soon] - saying “if AI could do this… then it could do this - but it cant - but it might” doesn’t seem to really counter my point that the next 5 years being full of empty promises about the potential (but not actualization) of AI.
I was more just giving context about how things work from a game dev perspective
On your first point, no, an individual AI is not, and never will be, capable of doing all of those things. What is will be is an analog to how the human brain works. You don’t see, hear, move, and process the words of a conversation you have while walking down the street with a friend using the same pieces of the brain. The occipital lobe, auditory and locomotive sections of the somatosensory cortex, and language center of the prefrontal lobe handle the parts independently of each other, then the information is brought together and presented to your conscious mind. An AI-driven questing system would have multiple specialized AIs that worked together to generate it. So a model which analyzes the current state of the player to determine valid reward thresholds and quest objective difficulties, another one which maps the current world lore to make sure that the quest fits into the world state, another which fills in all of the dialog based on NPC background variables, then a final AI which is trained to look at the outputs of the others to resolve conflicts. Finally, an AI voice synthesis can round out the experience for players. All of those can run in parallel and can use quite a few metrics from player interaction as feedback for refining the training.
To your second point, most of the aspects of a quest are rather small and can be stored in memory. The rewards can get interesting. If they are a world object, procedural modeling can go a long way to making it so asset generation is not necessary. If it is perks or traits of some variety, this could be something generational which uses keyword detailing to create the parameters for the trait. Generation and storage of details and items are not really much of an issue.
As for the engine questions, all of them can process geometry, textures, and text from memory or new files on disk. If something needs to be compiled, then it can be compiled on the fly. Again, individual assets are pretty lightweight and would not require a lot of processessing.
Another speed-up would be to pregenerate details of the quests rather than attempting to do it all using a just-in-time implementation. The game could generate the parameters for the world for NPC’s in town when you load in, starting with the ones closest to the player position and progressively iterating over them in the direction of travel. All you need to do is have details ready for the “chat bot” portion of the interaction by the time the player is capable of reaching any given NPC. These are the boundaries of what is possible so not as heavy as generating the whole thing. Then the rest can be filled in while the player talks with them.
The biggest issue I see is continuity error hardening. Making sure that all of the NPC’s worlds are consistent with each other and nobody makes changes that break the world for other NPC’s. That is specifically what I am trying to work on.
I’ve been hearing promises of the human brain being replicated by a PC since the 1980s, so again, politely, I consider that hyperbole/marketing gumpf until we see a working model.
I don’t particularly care if “AI” means a single model, multiple models, or multiple models banded together to appear as a single model or vice versa.
I didn’t realize you can just chop and change models on the fly, but taking those and similar issues as read (or at least probably solvable with modern day tech) — that leaves us with your multiple-AIs with specific functions.
Now I’m not saying it’s theoretically impossible, but i am asking: will you have a working prototype that can be run on a consumer home pc in the next 5 years? Or, are you, as I am very keenly aware of, simply doing what I stated in my first post: being a tech start up promising eventual, incremental process as product features?
Because my experience is, to generate a flat image from a model takes at least 20 seconds, not to mention 3D models with collision, mesh, animations etc. And 20 seconds is considered a long loading screen by modern standards. Gamers expect entire cities/planets to load in their game in under a minute.
Are we even close to generating even an untextured room with a single untextured T-pose NPC with no cues, triggers or animations from AI? Or, would it simply be using a language model to obfuscate the current and standard process for generating those assets to the user, when actually it’s just loading them from RNG.
This is unnecessary. The environment and characters can all still be built and rigged using modern hand modeling and set design techniques. The get here would be the animations. All animations, and I mean all, are just a bunch of splines and interpolations in the engine. Curve fitting is something that AI’s have been getting trained to do for 50 years. It is a solved problem. What is not solved, and the reason we don’t see a ton of AI animation tools already is a lack of sufficiently large sample sets of good animations attached to a variety of rigs and a method of training the AI to be environmentally aware. I imagine the latter could be solved using computer vision techniques with virtual sensors like what is used for Crowds. It is actually a hypothetical problem that I am planning on tackling after my current project.
This is almost insulting to the entire field of AI development, not to mention the mathematical fields of Probability Theory and Statistics. While AI models do function on Game Theory and Probability Theory models and are probabilistic, they are most definitely not any form of RNG. If you have that mindset, I suggest you do some more reading on the topics in scholarly publications or textbooks. Skip the pop-sci articles and blog posts.
As to your questions on timing, it may be this year, it may be 20 years. 2 years ago the idea of making photoreal images with AI was a pipe dream, then along comes Stable Diffusion and in less than 18 months we have gone from making passable images of a cat to an almost fully art-directable toolset capable of creating coherent videos. This is an amazing progressive leap and AIs in general have become more diverse because of it. It is a testament to just how fast something like this can grow given the right FOSS architecture and public interest. My guess is that it is closer to the 2 to 3-year range for a coherent world-building AI. That is not modeling, rigging, or animating, just the textual stuff. Story, relationships, lore, history, etc. My first tool I am trying to build is a world-building assistant for TTRPG GMs if that gives you a clearer picture of what I’m talking about.
Finally, the geometry generation times. You mention your experience, but I am struggling to pin down exactly what that entails for you. I mentioned using procedural modeling and having an AI that decides the parameters of the procedural when doing the modeling. In this setup the modeling is all mathematics and is done instantly. It can even include procedurally defined animations and affects that are able to be generated on the fly in microseconds. If you have 3D experience, I would suggest checking out Houdini ( https://www.sidefx.com ) to get a better grip on what is possible with proceduralism. They also have tools for doing rigging and animation that an AI could directly interface with, which can be utilized in game development realms as well as VFX.
You misread me when i said RNG - I was saying using a language model to appear as if it’s generating a quest on the fly, but actually its just picking Kill [5-30] [Goblins / Demons / Beasts / Mechs] from a list is not “AI generated quests”
And you misunderstood what I was saying about being a Dev - I said I was not a dev.
I know the following style of reply is a bit rude - but I want to highlight how I said that much of what tech companies are promising with AI is an empty promise that will only lead to semantic satiation of the term “AI” leading it to become a term of derision
so not AI.
so AI can’t do it (yet)
so AI can’t do it (possibly ever)
so unlikely to be a reliable marketing promise to consumers in this generation or the next
sounds like marketing bumpf
so not AI generated quests in game as originally promised…
Look, I really don’t mean to pick a fight with you - but saying “I’m working on AI generated quests!” to mean “I’m working on using an existing LLM to create text-based lore entries on my world” is a very different expectation to reality ratio.
Your response style is not rude. It is disingenuous confirmation hunting cherry picking. I addressed everything in your response in detail, but you latched on to the qualifiers and clarifications because those are the parts that satisfy your confirmation bias, even though in the full context they mean quite the opposite. I also am recognizing that you are not a dev, not a game designer, not an AI architect nor mathematician, or a computer scientist. All of this means that the scope and breadth of your understanding of the topic you are attempting to belittle and demean is myopic at best. It is obvious that you do not understand what a “quest” in a game is nor what it takes to craft or write one. It is clear that you have 0 understanding of how LLM or AIs as a larger topic function or generate information. You deign to belittle the work that I am doing on this topic without asking a single question or clarifying a single detail from me. I’m sure you are getting a nosebleed for your perch on Mt. Dunning-Kruger. I can see you from my position on the adjacent slope. I will humor you though, since you seem to at least care about the topic, even though you seem utterly incapable of recognizing when something should be informative and educational.
What is a “quest” in a game: The abstract of this article covers some key concepts of both current anf potential future paradigms of quest design in video games. Currently, quests, even handcrafted ones, consist of a list of tasks. Lists of tasks are text. Well-crafted ones include lore hooks, personal interactions with a variety of NPCs, tasks that are meaningful to the world, and rewards/results that either expand the player’s capabilities or draw the player into the world and the story. Components needed to craft a quest are a means of initiating the quest for the player (quest givers, discovered information like a letter or journal, an event that happens in the player’s presence), dialog, goals/tasks, and a location. For a well-made persistent world, locations, and NPCs are persistent as well. They don’t need to be AI-crafted on the fly. They should have been created when the game was made, or at least some period long before the players interacted with them. What can be crafted as a just-in-time or an emergent experience is the interaction with the player. Dialog with lore hooks that give the quest grounded footing in the world. Interactions with other NPCs which are based with the relationships of the NPCs with each other instead of just the player, etc. All of this is textual in nature. Dialog, relationships, interaction. The secondary stuff, like animations, could be handled in a lot of clever ways. They could be hand-crafted by animators and an AI could be trained on what animation links to what interaction type. The AI then selects and blends the canned animations when it is interacting with the player. And before you have something to say about that, it is literally how animation is done now by people. I absolutely have authority on this topic as I have a graduate degree in animation and visual effects. The AI I am working to create currently is being designed around the goal of making a consistent world with complex relationships and personalities so quests can be written that are self-consistent and coherent with the world. It is not worried about writing dialog or generating rewards, just making a world and quests that are cogent and have minimal plot holes.
What isn’t part of a quest? Geometry, models, textures, particles, VFX, rooms, people, trees, animals… basically anything that has physical form or defines a physical characteristic of an object. These things may be involved in the completion of the tasks themselves, but they are all separate from the quest. As an analog, your wife hands you a grocery list and asks you to go to the store to pick up the things she needs for dinner. This would constitute your bog-standard fetch quest in a video game. The quest is to obtain the listed items. All of the items exist independently of the quest. The grocer and cashier at the till do as well. As does your wife. The only thing that the quest consists of is the request which is made, and the list which is requested. Everything else does not need to be created, it already exists. I hope that clears up your misunderstanding of the topic.
You can erase the entire concept of modeling or texturing from your mind in relation to an AI creating quests. I mentioned it because I recognise that some objects may need to be created. Perhaps your wife asks for a product which the store did not contain already. So in a gaming context, this should be cleaned up so the quest is able to be completed. The product needs to be created. You can place limits on the list of items so they are all canned goods. Now you have a rudimentary prototype object that can be anything. Canned beans, sure. Canned orc tongue, why not? All that is needed is a label and a spot on a shelf. SD can create a label in a few seconds while the player moves from his home to the store and an observation-aware procedural system can stock it on the shelves when no players are looking and inform the AI driving the cashier of the existence and location of the item. Nothing needed modeled. Hell, all that was really needed was a procedural text generator to make a label for the can, no AI was needed for that. What the AI is needed for to make the trip to the grocery feel right is the request from your wife, the conversation with the cashier at checkout, and the smile and thanks from your wife when you got home with the items.
If more modeling is needed, procedural modeling takes care of that. Artists create prototype objects with fixed bounds on parameters and an AI is trained on how to set those parameters, so whenever a quest needs an item, the quest generation AI requests the type if item and the context in which it will be used, and the modeling AI interprets the request, prepares the correct prototype, determines the contextually appropriate parameters for it, then places it where the quest giver needs it. No muss, no fuss. Eventually, there will be SD analogs for 3D geometry and generative modeling. It is a field of active research at IBM, nVidia, AMD, Meta, and many others. It will happen, and it will happen sooner than anyone will be ready for it, much like SD. But for now, it is overkill. AIs can use procedural modeling to adequately furnish and populate a world without much overhead. I think I may play with that as well. Rip the character creator from the likes of Skyrim or Starfield and train an AI on faces and the character creator parameterization and have it go wild. The AI is still creating the characters, it is just using the same tool as the player to do so.
There are already mods that add voices to mod-scripted lines for Skyrim and Fallout 4. As well as a joke mod where the author took all of the recorded voice lines for Deep Rock Galactic(DRG), ran them through AI translators 40 times, then had an AI record the end result using the intonation and inflection of the characters in DRG.
However, the quotes from MS in the article provide an insight into their plans.
Interestingly, both of those show a fundamental lack of understanding of what a LLM can do. Yes it can do basic translation, but it fails on context in translations. E.g:
The French translation “you are a fool” can be “tu es un imbecile” but can also be “vous etes un imbecile” depending on the relationship context of the people talking.
And asking AI to replace QA entirely… oof, I guess I know to avoid MS games at launch from now on. Will be a lot of bugs.
I’m actually in the process of trying to get this setup to try myself. Wish me luck!