There’s nothing to complain about here. Games require tons of placeholders, in art, dialogue, and code. They will iterate dozens of times before the final product, and given Larian’s own production standards, there’s no chance anything but the most inconsequential or forgotten items made by an LLM will stay in.
Among the devs responding is a former Larian staffer, environment artist Selena Tobin. “consider my feedback: i loved working at @larianstudios.com until AI,” Tobin writes. “reconsider and change your direction, like, yesterday. show your employees some respect. they are world-class & do not need AI assistance to come up with amazing ideas.”
there’s no chance anything but the most inconsequential or forgotten items made by an LLM will stay in.
Concept art is not a placeholder. It’s part of the creative process. By using AI to generate text and images you already influenced the creative process negatively.
The article doesn’t say Larian is using it for concept art.
Those were hypothetical statements from people outside the studio.
Literally the first sentence
A few hours ago, reports surfaced that Larian are making use of generative AI during development of their new RPG Divinity - specifically, to come up with ideas, produce placeholder text, develop concept art, and create materials for PowerPoint presentations.
The actual quote from Swen is that they use it in the “ideation” phase of concept art. Basically: throwing shit on the wall and seeing what sticks. After that, the process is taken over by any of their almost 30 concept artists on payroll.
Yes, So AI comes up with the basic concepts.
No, it doesn’t.
Doodles are not concept art. Ideation is not concept art.
You’re right, I missed that
Idk but that seems pretty obvious to me from reading the quote by Larian CEO Swen Vincke that they used to, or are still using it to generate or “enhance” concept art and that’s it’s a highly discussed topic within the company?
What he actually said was that they use it in the “ideation” phase of concept art. Basically: throwing shit on the wall and seeing what sticks. After that, the process is taken over by any of their almost 30 concept artists on payroll.
Issue with non-obvious placeholder art is that it’ll easily fall through the cracks.
It’s not like they are just going to do a visual spot check on each level to clear out the AI assets. They will probably tag it in the meta data as a placeholder. So some automated validation process can find every ai asset in a level. Not to mention game objects are wrapped in an object template. And then the template is used to place the object into the scenes. So they only have to replace the placeholder with the final object once in the template and then it will replace it everywhere the template is used.
They honestly should have expected this given peoples visceral reaction to anything AI. Personally, I have huge problems with AI and refuse to play most games that have used it. I think it’s poisoning every creative industry and replacing important jobs while using vague the excuse that it makes things “easier” while making the game soulless in the process. I’m willing to give Larian the benefit of the doubt simply because of their previous games being amazing, but imma wait for the reviews on this one. This game is still going to be in development for another 4 years and none of us will no what’ll happen between then and now, but for now I’ll remain hopefully optimistic
Such a nuanced, unique opinion you have
Yeah, the outrage is overblown.
This doesn’t mean they’re enforcing a CoPilot quota or vibe coding the game or shipping slop; it could be simple autocompletion, or (say) a component that makes the mocap pipeline easier.
Don’t let Tech Bros poison dumb tools that could help out devs like Larian.
…Now, if they ship slop into the final game or announce an “OpenAI partnership,” that’s a different story.
…Now, if they ship slop into the final game
At a certain level, it is going to be a chore to determine who is or is not slopping up with AI media. Not every asset comes out with six fingers and a half-melted face.
I can see legitimate frustration with an industry that seems reliant on increasingly generic and interchangeable assets. AI just becomes the next iteration of this problem. You’ve expanded the warehouse of prefab images, but you’re still stuck with end products that are uncannily similar to everything else on the market.
And that’s before you get to the IP implications of farming all your content out to a third party that doesn’t seem to care where its base library is populated from.
At a certain level, it is going to be a chore to determine who is or is not slopping up with AI media. Not every asset comes out with six fingers and a half-melted face.
Image/video diffusion is a tiny subset of genAI. I’d bet nothing purely autogenerated makes it into a game.
I can see legitimate frustration with an industry that seems reliant on increasingly generic and interchangeable assets. AI just becomes the next iteration of this problem. You’ve expanded the warehouse of prefab images, but you’re still stuck with end products that are uncannily similar to everything else on the market.
See above. And in many spaces, there are a sea of models to choose from, and an easy ability to tune them to whatever style you want.
And that’s before you get to the IP implications of farming all your content out to a third party that doesn’t seem to care where its base library is populated from.
Thier tools can be totally in house, disconnected from the outside web, if they wish. They might just be a part of the pipeline on their graphics workstations.
Keep a distinction between “some machine learning in tedious parts of our workflows” and “a partnership with Big Tech APIs.” Those are totally different things.
It sounds like Larian is talking about the former, and I’m not worried about any loss of creativity from that.
At a certain level, it is going to be a chore to determine who is or is not slopping up with AI media. Not every asset comes out with six fingers and a half-melted face.
That’s a good point. I think a lot of people are dismissing AI content because there’s this fallacy and desire to believe it’s all “slop”. It’s willfully ignorant to wave it all away like that. Sure, we’ve all seen the stupid stuff, and it’s really annoying, but we absorb the good stuff without even knowing it. Anyone claiming they can reliably spot AI generated images is fooling themselves even at this early stage.
I’d like to know when something is real, especially real art and even pictures of nature, but I don’t think I can.
I’ve had an idea of making a visual novel with gen AI, but I’d want to attach “Placeholder: AI Artwork” in a visible location for each sprite. And I only even consider that because I’m not exactly a known game dev and don’t have ready access to artists.
Larian should likely expect if they’re taking shortcuts in their position, they’d get backlash. I can at least recognize that they’re trying to be moderate about it.
Also you don’t have the infamous AAA deadlines so you are not guaranteed to include them in final outcome anyways due to “an oopse.”
Larian has kept the plot.
I think they’re one of the few studios that actually appreciates the success they’ve been given and just want to make good games.
Welp, there goes all of my enthusiasm for the next game. Will have to check out other actual game developers and artists instead of whatever Larian and their genAI prompt monkeys have become.
We’re not looking to, but if it happens, it happens.
Gamers will and do bitch about anything and everything. I couldn’t care less if they use AI if the product is good.
Yeah I am sure higher ups of this one studio alone will use it responsibly because they are “not like the others” and definetely will resist making a bad decision despite deadlines.
I would play a game with ai NPCs. Not the artwork but the mechanic. You could run hundreds of interqctable characters with simple prompts that states ztheir goals and personalities. Like a modern radiant ai.
These games already exist and they’re just boring, no thought behind any dialogue, just a waterfall of LLM slop.
any name I can take a look at? I’m curious to see it.
I’m not sure about games personally, but there’s the Mantella mod for Skyrim which is pretty much exactly what you’re looking for.
Check out this video if you want to see how it works in practice.
Cool video that show the potential. Add different voices and vocabulary levels and it become quite interesting. Maiq would be great liar.
Why is one thing okay for you but not the other? A generated artwork is an artists not paid and theft. Generated dialogue is a writer not paid and also theft. All the popular models are fed on stolen content.
Valid questions. There are many issues to distinguish in your question.
AI is Theft, Popular models are fed stolen content. I agree. They should be sued for all they they have. Either pay or make them a world heritage, open to all without fees.Why NPC dialog and not models or asset? Assets and world building are designed one an usually fixed. Which make sense the world is a fixed space with fixed rules. Having AI generated scripts for NPC would fall in the same category and that is not what I would try. What I expect are au powered well crafted personages that live in that world. Rondo that you still need good writers, to create the persona background and goals and quirks to make it interesting to interact with. Your companion can only be sworn to carry your burden so many times. Charisma and factions stats become way more engaging etc. AI is also not limited to LLM and generative, having a faction figuring out that some sneaky bastard is robbing them could trigger them to plant honey pots to try and catch you. Ennemies learning would be cool
Games dead on arrival clanker.
“But free assets are free assets.”
They’re not using AI for free assets.
“Pre-visualization for concept art” is just a fancy way of saying “we want to use Keith Tompson’s designs, but we don’t want to pay his asking price.”
Yeah, you clearly have no clue what you’re talking about, but just want to be outraged.
No, that’s not what they’re using it for.
Pre-concept art phase is when you grab literally anything to do a very generic “here’s my idea” demonstration. It can be a screenshot from a game you recently played, a cinema poster, a photo of your cat, something you found on DeviantArt or Pinterest, a doodle you made with a pencil, cloth fragment, an interesting rock you found on a stroll, simple render, anything.
But, getting all these takes a lot of time - you have something in your head and now you need to find an image or an item that will more or less represent it. So you spend hours on Google Images trying to refine your search, only so that you can then post it on the ideas board, and for it to be replaced completely by actual concept art.
This is where they’re utilising GenAI. And they’re not even replacing this process entirely, they’re using GenAI on top of everything else - basically, using all the tools available to speed up the process.
Concept artists then still take over (they have 27 of them and open positions to hire more - right now) and create concept art, which is then turned into assets by appropriate artists (they have a bunch of open positions for artists right now as well).
All this anti-AI panic has blown this so far out of proportion that it’s almost comical.
I really hope this marks the end of the internet’s weird love fest with this company and CEO. They don’t care about you in the slightest.
I dont give a shit if they care or not, they make amazing games
