• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Antigonid1

Corporal
Sep 27, 2022
47
339
Could ChatGPT or other AI be the solution to the questionable quality of events ? I believe the AI could easily handle the code behind the events. Let's face it, it would be a vast improvement over the farts jokes and general lack of immersion provided by the writing
 
  • 71
  • 8Haha
  • 6Like
  • 1
Reactions:
Here is a dev response about this very same topic. Hope it helps explain why devs think it is a bad idea :)

P.S.: there are also other devs' thoughts in that same thread, I just linked the first one.
 
  • 14
  • 6Like
  • 3
  • 1
Reactions:
Could ChatGPT or other AI be the solution to the questionable quality of events ? I believe the AI could easily handle the code behind the events. Let's face it, it would be a vast improvement over the farts jokes and general lack of immersion provided by the writing
Oh, absolutely! Because nothing says "immersive event experience" quite like surrendering to our AI overlords. Let's just hope the code doesn't include fart jokes in binary—now that would be a real gas!

(written by ChatGPT)
 
  • 15Haha
Reactions:
I … there’s a second thread now about using a chatbot to write events? Is someone using a chatbot to write these?
90347968-4C1F-4458-AD2F-C3C6630041D6.jpeg
 
  • 26Haha
  • 1Like
  • 1
Reactions:
There are already books written by AI. I won't be surprised if in 5 years most quests/events in video games will be designed by AI.
 
  • 13
  • 2Like
  • 1
Reactions:
I won't be surprised if in 5 years most quests/events in video games will be designed by AI.
Neither would I, but that's mostly because corporations (a) are skinflints (b) know a lot of players are going to skip through all the quest dialogue anyway.
 
  • 15
  • 2Like
  • 1Love
  • 1Haha
Reactions:
There are already books written by AI. I won't be surprised if in 5 years most quests/events in video games will be designed by AI.

My only hesitation here is that many video games use unique syntax that some sort of big data model like these language models will struggle to grasp simply because there has not been enough of it written (by people) to train the model, barring the fact even that most of it is closed source. Language was an easier problem because written language is so common and has been written and published for a hundred years* with an accelerating rate. This given, writers for plots and quests may dramatically decrease in number. This will be a shame since chatGTP writing still isn't on par for an actually good writer in my opinion.

*use things much older than a hundred or two hundred years and you might have the model learning archaic things it shouldn't
 
Last edited:
  • 2
Reactions:
There are already books written by AI. I won't be surprised if in 5 years most quests/events in video games will be designed by AI.
Well it certainly could and I wouldn't be surprised either, but there is a major difference between making a videogame which uses AI scripting/writing from the start and adding it later, which is why I respectfully disagreed to OP.

I mean, other reasons as well but I do not wish to engage with this topic :)
 
  • 1Like
Reactions:
:) Steam Workshop's available, loc files are there to train it, ChatGPT is free, welcome to give it a bash!

I'm honestly expecting someone to start managing a curated mod of AI-generated human-fixed content at some point. I'm rather sceptical about the potential quality, honestly, but I'm hardly unbiased on that front. I do wonder how folks intend to combat the loc bias, because unless you manually retrain it on this or that mod (preferrably with a modder's permission), the script & event types it's going to be imitating will presumably be based on our stuff anyway, even if you tell it to "script and write an event for CK3 in the style of [favourite author]".

That said, to my mind, the main concern is that the essential proposition is to turn event creation into 100% debugging and editing. The bits that take the absolute longest to do, are the least fun, and which require the most effort. It does not actually take that long to bang out a concept, an event, and the loc for it - it takes markedly longer to make sure that it's not causing errors which might affect stability or cause other bugs, to play-test it for interesting effects, to check that the loc flows smoothly in-game, and of course, to account for edge cases (which range from the more simple stuff like "what if a character is blind, has one arm, one leg, or doesn't technically speak your language?" all the way through to the advanced fields of "what if they're currently on holiday in Jerusalem right now and I've accidentally referenced them standing next to me at a party?" or "what do I do if I a character has an obscure cultural tradition that means they should only do The Thing Wot The Event Is About when the moon is full?").

All of this is much easier on your own stuff than someone else's stuff, because you know what you were thinking and what individual parts mean or what the intent was, whereas for someone else, you need to guess, intuit, figure it out, or rely on their script comments (if they left any, which they bloody well should have). Already not a great start for editing an AI's content to be useable, except it's sorta worse than you might think, because the AI doesn't have an intent. It's not coordinating which bits go together and which don't, or thinking about any of these things. It's just aping what it sees in a semi-intelligent fashion, but doesn't understand why it's doing any of this stuff. So, now add on to the debugging/editing process that you need to retroactively construct an intent for the event, probably working outwards from the prompt you gave the AI or its initial response, and edit or amend potentially anywhere between just a few things and literally everything it wrote to account for that.

If there's events swirling around your mind right now as examples of things that don't meet this standard, then yeah, behold examples where we probably should've spent more time on editing & debugging, or where some of this stuff wasn't accounted for early enough or thoroughly enough. An AI isn't accounting for anything at all, so its baseline is far, far below this - and this is before we get into more ephemeral concerns like the practicalities of event chains (which go up semi-quadrilaterally with however many extra events/permutations you put in that need to account for what may or may not have happened between events), whether or not it's repeating a concept that's already been done in the title before in some fashion, whether it's violating design imperatives (which are sometimes just tonal and totally something you can cross with a mod, and sometimes fundamental to how the title is expected to function but not actually stated anywhere and breaking them will just screw the entire game up), or, of course, historicity (especially operating in a period with a lot fewer sources - most especially textual sources - than something like V3 or HoI4).

You could probably do it, even so, but the amount of overhead created by having an AI make something that likely doesn't even initially work and also still needs a veteran scripter & writer to repair whatever it was trying to do massively exceeds the gain compared to just having that designer do their job ordinarily.

I do think it'd be kind of interesting to see an AI-created event mod for CK3, I know I'd play it for a bit at least, but I definitely don't fancy the actual work needed to produce one.
 
  • 28
  • 15
  • 2Like
  • 2Love
  • 1
Reactions:
Well, it's definitely a sign for developers to think about their approach to events when there are people who think that a chatbot would do a better job of writing events than the developers themselves. :rolleyes:

But no, I'm really not a fan of such ideas. I still prefer a creative approach to writing, even if that creativity is often very poorly used.
 
  • 21
  • 3
  • 1Like
  • 1Haha
Reactions:
Well, it's definitely a sign for developers to think about their approach to events when there are people who think that a chatbot would do a better job of writing events than the developers themselves. :rolleyes:

But no, I'm really not a fan of such ideas. I still prefer a creative approach to writing, even if that creativity is often very poorly used.
It's not really a sign because anybody who thinks an ai could write events would re-write the Lord of the rings with an ai if they could.
 
  • 23
  • 1Like
Reactions:
It's not really a sign because anybody who thinks an ai could write events would re-write the Lord of the rings with an ai if they could.
That's why I wrote that I'm not a fan of such ideas and prefer a creative approach to writing. :)

That first part of my post was a joke. But that joke was trying to make a point. A lot of players obviously feel that something is wrong with events and they are trying to come up with a solution, sometimes this extreme to have the AI generate them.
 
  • 4
  • 2
Reactions:
That's why I wrote that I'm not a fan of such ideas and prefer a creative approach to writing. :)

That first part of my post was a joke. But that joke was trying to make a point. A lot of players obviously feel that something is wrong with events and they are trying to come up with a solution, sometimes this extreme to have the AI generate them.

A lot of players also massively overreact to things and then get really mad when their specific issues aren't dealt with immediately. Just because someone jumps to the extreme of "let an AI bot write events" doesn't mean the problem that suggestion is supposed to be a solution for is actually that much of an issue.
 
  • 13
  • 1Like
Reactions:
That's why I wrote that I'm not a fan of such ideas and prefer a creative approach to writing. :)

That first part of my post was a joke. But that joke was trying to make a point. A lot of players obviously feel that something is wrong with events and they are trying to come up with a solution, sometimes this extreme to have the AI generate them.

Players tend to talk a lot about the same events 'cat a pult' or 'farts' while ignoring hundreds or thousands of events that don't bother them (except when they're repeating, which is another debate). I played 2 hours last night and none of the events I encountered were weird or badly written (at least in French, maybe our translators are Goncourt Price winners, but I doubt it).
 
  • 9
  • 2Like
  • 1
Reactions:
Players tend to talk a lot about the same events 'cat a pult' or 'farts' while ignoring hundreds or thousands of events that don't bother them (except when they're repeating, which is another debate). I played 2 hours last night and none of the events I encountered were weird or badly written (at least in French, maybe our translators are Goncourt Price winners, but I doubt it).
Exactly, there are 3-4 stupid events that are decently rare (I think one of the devs said they had statistics and that most of the people complaining about the cat-apult most likely haven't even seen it) and because of this some people complain about how CK3 is just the worst game ever made.
 
  • 8
  • 3Like
  • 2
  • 1
Reactions:
Players tend to talk a lot about the same events 'cat a pult' or 'farts' while ignoring hundreds or thousands of events that don't bother them (except when they're repeating, which is another debate). I played 2 hours last night and none of the events I encountered were weird or badly written (at least in French, maybe our translators are Goncourt Price winners, but I doubt it).
As many of us have pointed out many times, the cat-a-pult (and farts and similar events) are more a symbol of what bothers us. It's not a criticism of that particular event, that event is a personification of broader issues. There are many of those issues, and bad writing (more specifically the unbelievability of the events) is just one of them (not to repeat the same rare events over again, a tinder event or a very common birthday event are also very out of a place and immersion breaking). Repetition is another one of the problems, because even a good and interesting event, if repeated very often, starts to get boring and annoying (for example a headless saint event). But as I've realized, the repetition is just a symptom of a deeper problem, and that is a strong focus on events. Instead of CK3 being a sandbox game where the story is generated by the mechanics and imagination and the events are the flourish on top of it (like in CK2), in CK3 the events are the main component of the game. This is very limiting because events are pre-made stories that we get over and over and over and over again in the same form... And despite the developers claiming to focus on roleplaying, such a strong focus on events limits roleplaying. Because instead of creating my own story while playing the game, I'm fed stories that the developers have pre-created. Playing CK3, I feel like I'm experiencing the same stories over and over again, with slight differences. The developers are trying to diversify it by introducing new events, but that's only a temporary solution.

Edit:
Exactly, there are 3-4 stupid events that are decently rare (I think one of the devs said they had statistics and that most of the people complaining about the cat-apult most likely haven't even seen it) and because of this some people complain about how CK3 is just the worst game ever made.
No one is saying CK3 is the worst game ever made. All of us here love CK, otherwise we wouldn't be here playing it. But that game has its problems. Even you admit that some of the events are stupid. While all the developers here have always defended even events like cat-a-pult, saying it's a slightly amusing event and they don't understand what the problem is. Instead of acknowledging that they pushed it a bit too far. This is a big red flag to me, because it makes me feel like turning the game into a meme is their intention.
 
Last edited:
  • 9
  • 8
  • 2
  • 1Like
Reactions: