When Larian Studios speaks, the industry listens and not just because of Baldur’s Gate 3 or the newly announced Divinity project. Larian has become a rare example of a large studio that pairs scale with strong creative values. That’s why CEO Swen Vincke’s comments about AI use at The Game Awards 2025 sparked such a strong reaction.
Vincke was clear: Larian uses AI internally to explore ideas, sketch rough concepts, draft placeholder text, and support early ideation. None of that material appears in the final game. And yet, social media responded as if Larian had crossed a creative red line. The controversy reveals less about Larian’s practices and more about the industry’s anxiety around where AI belongs in game development especially when authorship, immersion, and trust are at stake.
AI as Exploration, Not Output
Larian’s position is straightforward. AI is used at the very beginning of the creative process, where ideas are fluid and disposable. Rough compositions, reference exploration, and placeholder drafts help teams move faster but they are always replaced by human-authored work. Vincke was explicit: Larian employs 72 artists, including 23 concept artists, and is hiring more.
This distinction matters. There is a profound difference between AI as a thinking aid and AI as a content generator. Larian’s pipeline treats AI the same way studios have long treated mood boards, reference books, or Google Image Search as a starting point, not a deliverable.
AI as Tool vs AI as Author
| Usage Layer | Role of AI | Creative Ownership |
|---|---|---|
| Early ideation | Rough exploration | Human-led |
| Reference gathering | Visual inspiration | Human interpretation |
| Placeholder drafts | Temporary scaffolding | Rewritten by humans |
| Final assets | None | Fully authored |
Insider Tip: If AI output survives into the final build, it’s no longer a tool, it’s an author. Draw the line early and enforce it strictly.
Why the Backlash Happened Anyway
Despite Larian’s clarity, the backlash was immediate. That reaction didn’t come from misunderstanding it came from fear. Across the industry, AI is being aggressively pitched as a replacement for creative labour, especially in concept art, writing, and narrative design. Developers have watched colleagues laid off while executives talk about “efficiency.”
In that context, any mention of AI triggers alarm bells. Even responsible use is viewed with suspicion. Larian became a lightning rod not because it did something wrong, but because trust in how AI is being deployed elsewhere has already collapsed.
Why AI Discussions Escalate So Quickly
| Industry Context | Developer Perception | Result |
|---|---|---|
| Layoffs framed as automation | AI equals job loss | Immediate hostility |
| Vague corporate language | Lack of trust | Assumed bad faith |
| Player-facing AI experiments | Immersion damage | Strong resistance |
Insider Tip: Transparency isn’t optional anymore. Studios must over-communicate AI boundaries or risk being lumped in with bad actors.
Authorship Is an Immersion Problem, Not a Moral One
From an immersive design perspective, the issue isn’t whether AI is ethical. It’s whether AI undermines authorship. Immersion depends on coherence: tone, intent, voice, and meaning must align across systems, narrative, and world design. Generative AI struggles with this because it produces plausible output without understanding context.
Larian avoids this problem by ensuring AI never touches the player-facing layer. Narrative, characters, environments, and systems are all authored by humans who understand the thematic and systemic logic of the world. That’s why Larian’s games feel intentional, consistent, and deeply immersive.
AI Placement and Immersion Risk
| AI Placement | Immersion Impact |
|---|---|
| Internal ideation | None |
| Production support | Low |
| Narrative generation | High |
| Player-facing dialogue | Severe |
Insider Tip: Immersion breaks the moment the player senses a lack of intention. AI has no intention – only probability.
Larian’s Approach as a Model for the Industry
What Larian demonstrates is not anti-AI sentiment, but disciplined use. AI is treated as middleware, infrastructure that supports creativity rather than replacing it. This is the model most studios should adopt if they want to remain resilient in an unstable AI landscape.
External AI models change rapidly. Licensing terms shift. Public sentiment fluctuates. Studios that rely on AI-generated final content expose themselves to legal, creative, and reputational risk. Studios that keep AI behind the curtain retain control.
Middleware AI vs Player-Facing AI
| Approach | Risk Profile | Long-Term Viability |
|---|---|---|
| Internal tools | Low | High |
| External content generation | High | Unstable |
| Human-authored output | Low | Durable |
Insider Tip: AI should accelerate the process, not define the product.
Final Thought
Larian’s defence of its AI use is important not because it justifies the technology, but because it draws a clear boundary around authorship. AI is not creating worlds at Larian. Humans are. AI is simply helping them think faster, explore broader references, and reduce friction during ideation.
The games industry doesn’t need to reject AI outright. It needs to put it in its place. Tools belong behind the scenes. Meaning belongs in human hands. Larian understands this and that’s why their games continue to feel authored, intentional, and alive. The future of immersive design will not be written by machines. It will be written by people who know how and when to use them.
That’s it for this one! Please like, share, and comment if enjoyed this article AND…
Grab Myles’ FREE ebook now and find 15 indispensable design patterns that will equip you to craft exceptional Web3 gaming experiences. We’ll also notify you when his new book on immersive design is out!