Okay, so I was reading this whole Reddit thread about Larian Studios—you know, the folks behind Baldur’s Gate 3—and their CEO, Swen, basically had to clarify their use of AI in concept art. It all started because he mentioned they use generative AI to “explore things,” and, well, the internet did what it does best: exploded. His clarification post even opened with “Holy fuck guys,” which a lot of people found refreshingly human, like a CEO actually talking like a normal person and not a corporate robot. But that’s where the agreement sort of… splits.
On one side, you’ve got folks who really appreciate Larian’s transparency and see their use of AI as totally sensible. The analogy that got a lot of traction was that they’re using AI to “throw shit at the wall to figure out where the wall is,” and then actual, world-class artists step in to create the real concept art. It’s like using Google Images or art books for reference—just a tool in the early, messy ideation stage to spark ideas or help communicate a vision when words fail. They’re not replacing their team of concept artists; they’re hiring more. For these people, it’s a pragmatic, modern workflow, and the backlash feels like an overreaction from a crowd that hears “AI” and immediately panics without nuance.
But then, the other side is deeply, fundamentally worried. And it’s not really about Larian specifically—it’s about the slippery slope. The fear is that today it’s “just for exploration,” but tomorrow it becomes “just for concept art,” and then it seeps into everything. There’s a strong undercurrent of concern about AI normalizing job displacement in creative fields, even if Larian isn’t doing that now. People point out that the “exploration” phase *is* a core part of a concept artist’s job, and outsourcing even that to a machine chips away at the creative process. Plus, there are bigger, systemic issues tied to AI that folks are angry about: the environmental cost of massive data centers driving up electricity bills, the way it’s inflating hardware prices like GPUs and RAM, and the fact that the underlying models are often trained on artists’ work without permission or compensation. For this group, any use of generative AI, however minor, feels like endorsing a harmful and predatory system. They want regulation first, before any “nuanced” discussion about it as a tool.
What’s really fascinating, though, is how the discussion spiraled into this meta-debate about gamers and developers talking to each other. A lot of commenters sighed and said, “This is exactly why most companies don’t engage openly anymore.” They pointed out that developers are damned if they do, damned if they don’t. Gamers, as a demographic, can be notoriously harsh and bad at proposing solutions, even if they’re good at spotting problems. Some said Larian’s more refined fanbase might exhaust itself with internal debates before the vitriol reaches the devs, but others cited examples like Old School Runescape or Guilty Gear Strive, where devs learned to listen to feedback to identify issues but then rely on their own data for solutions. The consensus seemed to be that the sweet spot is to listen, analyze, but not try to make everyone happy—because that’s impossible.
There was also this interesting layer of… almost cynicism about trust. Some said, “I trust Larian because they make great games,” but others immediately pushed back, naming companies like Rockstar or old Blizzard that made amazing games but were scummy behind the scenes. The more cautious view was, “Don’t *trust* any company—they’re profit-driven entities. You can *expect* responsible behavior from Larian based on their track record, but blind trust is naive.” This tied into a point a few made about hypocrisy: if a disliked company like EA or Ubisoft had a CEO post “Holy fuck guys,” they’d be ripped apart for being unprofessional. It’s a double standard based on goodwill.
A few specific, smaller threads stood out. One person linked to a claim from an alleged ex-Larian employee about AI being pushed on them, which added fuel to the “don’t trust corporations” fire, though others questioned the source. Another really detailed comment argued that the core issue is a lack of legal regulation, making it impossible to have a grey-area discussion—everything is black or white until laws catch up. And there was a pragmatic observation that this kind of AI use for mood-boarding is probably already widespread in the industry; Larian is just getting heat for being transparent about it.
So, in the end, the thread was less about Larian’s specific practices and more a microcosm of the huge, messy AI debate. You have the pragmatists who see a useful tool, the idealists who see an existential threat to art and jobs, and the realists in the middle who are tired of the lack of nuance but understand why emotions are running so high. All sparked because a CEO used a casual “Holy fuck” to defend his artists. The internet, man.
