Editorial illustration for House House says LLM snippets used in Big Walk code amid indie AI dispute
House House Used LLM Code in Big Walk Game
House House, the Australian studio behind Untitled Goose Game, spent the past day clarifying exactly how much artificial intelligence went into its latest release, Big Walk. The admission landed in the middle of a fight that's splitting the indie game world down the middle.
The trouble started with a statement from a group of Independent Games Festival jurors, posted online by Austin Walker, declaring they won't vote for any game built with generative AI "in any capacity." Signatories include Sam Barlow, who made Telling Lies and Immortality, and Celeste composer Lena Raine. The jurors say their position doesn't represent the IGF itself, and some reportedly worried about putting their names to it at all, fearing "professional blowback" from the industry.
Big Walk became the flashpoint almost immediately. The co-op adventure game has been one of the festival's buzzier entries this year, and questions about its development process turned into a referendum on where the line sits between using a tool and outsourcing craft. House House's response, addressing exactly what role AI played in finishing the game, is below.
Some Independent Games Festival (IGF) jurors published a statement yesterday saying they would not vote for games that use generative AI in “any capacity.” Soon after, the developer of indie hit Big Walk has published a statement saying it “sparingly” used AI in the “final stretch” of the game’s development to “help work through difficult technical challenges and bugs.”
Why this matters
The IGF jurors' blanket stance, no AI "in any capacity," was always going to collide with how studios actually build games. House House's 0.1% figure shows the gap between that rule and practice: a small team under deadline pressure pasted LLM output into code to fix bugs, not to generate art or design. That's a far cry from the wholesale AI authorship critics seem to be picturing, but the policy as written doesn't distinguish between the two.
For developers, the lesson is that vague thresholds invite exactly this kind of dispute. "Any capacity" sounds principled until a beloved indie hit forces jurors to define what counts, and a one-line debugging assist starts getting weighed against full AI-generated assets. For festival organizers and award bodies watching this play out, the fix isn't more outrage, it's actual disclosure standards: percentage of code, what was AI-assisted versus AI-generated, and for what purpose. Without that, every studio using AI even minimally for troubleshooting risks getting lumped in with wholesale generation, and juries end up litigating intent instead of evaluating games.
Related Reading
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source Axiom Math raises USD 200M Series A led by Menlo to keep AI slop out of math
From Archives Agent Substrate Manages Stateful Agents Above Kubernetes at Scale
From Archives Hugging Face exec warns cyberattacks threaten open-source AI
Common Questions Answered
How much generative AI did House House use in Big Walk's development?
House House stated they used AI sparingly during the final stretch of Big Walk's development, specifically to help work through difficult technical challenges and bugs. The studio's admission revealed they used LLM snippets in approximately 0.1% of the game's code, representing a minimal amount of AI integration focused on problem-solving rather than creative generation.
What was the Independent Games Festival jurors' stance on generative AI in games?
A group of IGF jurors published a statement declaring they would not vote for any game built with generative AI "in any capacity," establishing a blanket policy against AI use regardless of how minimal or limited its application might be. This strict position created significant controversy within the indie game development community.
Why does the IGF's blanket AI policy create challenges for indie developers?
The IGF's "no AI in any capacity" rule doesn't distinguish between different types of AI usage, such as using LLM snippets to fix bugs versus wholesale AI authorship of art or design. House House's 0.1% figure demonstrates the gap between the policy as written and how small teams under deadline pressure actually use AI tools to solve technical problems, creating a collision between the rule and real-world development practices.
What specific purpose did House House use LLM snippets for in Big Walk?
House House used LLM snippets specifically to help work through difficult technical challenges and bugs during the final stretch of Big Walk's development. Rather than using AI for creative tasks like art generation or game design, the studio employed it as a practical tool for code troubleshooting and problem-solving.
How does House House's limited AI usage differ from the concerns critics have about AI in games?
House House's 0.1% AI usage in code snippets for bug fixes represents a far cry from the wholesale AI authorship of creative content that critics seem to be concerned about. The distinction shows that small teams may use AI minimally for technical assistance rather than for generating art, design, or other core creative elements of the game.
Further Reading
- Papers with Code - Latest NLP Research - Papers with Code
- Hugging Face Daily Papers - Hugging Face
- ArXiv CS.CL (Computation and Language) - ArXiv
