House House says Big Walk used LLMs sparingly for late-stage technical problem solving, prompting a wider debate over indie game AI tools, IGF voting, and player trust.
Image: IGDB
Store links: Big Walk on Steam
House House’s AI disclosure landed in the middle of an awards fight
House House has confirmed that Big Walk used large language models during the final stretch of development, a disclosure that immediately put one of the year’s most-discussed indie games at the center of a broader fight over generative AI in games.
In a statement first reported by Aftermath and later covered by GamesIndustry.biz, Nintendo Life, and Rock Paper Shotgun, the Australian indie studio said “LLMs were used sparingly by House House to help work through difficult technical challenges and bugs.” The developer estimated that “well under 0.1% of Big Walk’s code had anything to do with this usage” and added that “generative AI was not used in any other context when making the game.”
That last sentence is doing a lot of work. House House is drawing a line between limited chatbot-assisted code troubleshooting and the use of generative AI for art, writing, audio, design, or other creative assets. The studio is also acknowledging that, for many players and developers, that distinction will not settle the issue. The Big Walk AI use disclosure arrived after a group of Independent Games Festival jurors publicly said they would decline to vote for any submitted game that used generative AI “in any capacity.”
The result is a compact but revealing indie game AI tools controversy: a beloved-scale studio describing a tiny technical intervention, critics arguing that any use crosses a line, and players trying to decide whether “well under 0.1%” is meaningful context or beside the point.
What House House says actually happened
House House’s account is specific about both the scope and the kind of AI use involved. According to the developer’s statement, LLMs were used in the “final stretch” of Big Walk’s development to help research technical problems involving third-party code. In some cases, the studio said, snippets from LLM responses were copied as a starting point and then modified into game code.
The studio’s clearest example involves Unity’s Universal Render Pipeline. House House said it needed to cull shadow casters by distance from the camera, but Unity URP only allowed shadow casters to be culled by distance in camera forwards. In the studio’s explanation, that meant shadow casters behind the player were being rendered at an infinite distance, which made Big Walk “run 50% slower for around 20 seconds at sunrise and sunset.”
House House said it researched the issue, inspected URP code, and spoke directly with a Unity engineer, who asked around and ultimately concluded that the studio’s desired solution was not possible. The developer then posed the question to an LLM chatbot. “After a few dead ends, it suggested a possible fix, and while we had to modify the suggestion extensively, it worked,” House House said. “This fix represented about 10 lines of code inside URP.”
On the confirmed record, then, the House House AI statement is not a claim that Big Walk contains AI-generated art or AI-written narrative material. The developer explicitly says it does not. The confirmed claim is narrower: LLMs were consulted for late-stage technical debugging, and some chatbot-derived snippets were used as a starting point for a very small amount of code.
The timing is the story, not only the code count
The disclosure matters because it did not arrive in a vacuum. GamesIndustry.biz reports that Big Walk was submitted for the 2027 Independent Games Festival, with submissions closed on September 13 and finalists due to be announced in January 2027. House House said it wrote “a version of” the disclosure for the IGF, where AI disclosure was a suggested part of the game’s submission.
The public statement followed a post from a group of IGF jurors who said they had seen a “significant increase in the number of submitted games developed with generative AI tools.” Those jurors wrote that they would decline to vote “for any game that uses generative AI in any capacity” and urged the IGF to adopt a future policy disallowing games containing AI-generated work. Rock Paper Shotgun also notes that the juror statement specified that “AI-generated code is not an exception to these harms.”
That puts Big Walk in a difficult category. House House is not being accused, based on the cited reporting, of hiding AI-made creative assets in the game. At the same time, the IGF jurors’ position does not carve out a safe exception for code assistance, debugging, or a small percentage of a codebase. For those jurors, “any capacity” means the small technical use House House described still matters.
This is why the timing has become central to the Big Walk controversy. The studio did not put this disclosure on Steam, and House House says that is because Steam “does not request disclosure of code or chatbot usage.” Instead, the developer said it was posting the disclosure because “it has become a matter of public interest.” In other words, the pressure point was not a store compliance requirement. It was the collision between awards transparency, community scrutiny, and a changing norm around what players expect indie developers to disclose.
House House’s apology is careful, and also conflicted
House House’s statement does not frame the LLM use as a proud production breakthrough. The studio says, “This is not an endorsement of, or argument for, the use of AI tools, however minimal.” It also writes that it has “some concern that voluntary disclosures like this might contribute to the normalisation of this technology, or its treatment as ‘inevitable.’”
The developer goes further, saying it supports “the work of those who are trying to limit the harm generative AI is doing to videogames and our world” and apologises that Big Walk’s LLM usage “may not meet their standards.” House House also says Big Walk was developed “without an internal AI usage policy in place” and calls that lack of policy “a failure,” adding that it will work toward a more comprehensive policy going forward.
That is an unusually cautious statement for a developer trying to contextualize a small technical use. It accepts the premise that generative AI has harms, admits the studio lacked an internal framework, and avoids using the successful bug fix as a sales pitch for the tools. But it also leaves a hard question unresolved: if the studio shares concerns about normalizing generative AI, why did it use an LLM at all once internal research and a Unity contact had failed?
Rock Paper Shotgun’s coverage lands on that contradiction sharply, criticizing the decision as voluntary rather than imposed by a publisher or management mandate. That criticism reflects a wider discomfort in the indie scene. Small teams are often admired for visible authorship, deliberate craft, and human-scaled problem solving. When one of those teams says it used an LLM, even sparingly, the percentage can feel less important than the precedent.
Players are splitting over scope versus principle
Aftermath described the discussion around Big Walk as “deeply divisive,” with debate emerging across social media after the IGF jurors’ anti-AI stance became public. According to Aftermath, some people saw the news as a major disappointment despite the seemingly minor use. Others treated it as a sign of the times, or as a case where suspicion of AI tools may be running ahead of the specific facts.
That split is easy to understand. On one side, House House’s disclosure gives players concrete limits: the studio says generative AI was not used outside this technical context, the amount of affected code was well under 0.1%, and the cited fix amounted to about 10 lines inside URP after extensive modification. For players mainly worried about AI replacing artists, writers, musicians, level designers, or performers, the developer’s statement directly says that did not happen here.
On the other side, many players and developers do not treat generative AI as acceptable if it stays away from art assets. The IGF jurors’ statement, as cited by Rock Paper Shotgun, explicitly rejects an exception for AI-generated code. From that perspective, the ethical concern is tied to the tools themselves, their training practices, their labor implications, their environmental cost, and their normalization across production pipelines. Under that standard, a small amount of code assistance still counts.
For a co-op indie adventure like Big Walk, that tension is especially sharp because player affection for small games often comes from trust. People want to believe in the hands behind the strange walk cycle, the landscape, the social friction, the tiny technical flourishes that make an indie game feel authored. House House’s statement preserves many of those boundaries on paper, but the reaction shows that disclosure after the fact can still dent confidence.
The practical read for Big Walk players
For anyone deciding how to respond, the confirmed facts are narrower than the debate surrounding them. House House says Big Walk used LLM chatbots late in development for difficult technical challenges and bugs. It says some snippets from LLM responses were copied as starting points and modified. It estimates that well under 0.1% of the game’s code had anything to do with that usage. It says generative AI was not used in any other context when making the game.
GamesIndustry.biz lists Big Walk in connection with Mac, Nintendo Switch 2, PC, and PS5. The cited reporting does not provide new pricing, release-date, or performance guidance beyond House House’s technical example about the sunrise and sunset slowdown. It also does not announce an IGF ruling on Big Walk’s eligibility or finalist status. Finalists, according to GamesIndustry.biz’s reference to IGF submission information, are due in January 2027.
If any generative AI use is a personal deal breaker, House House has now confirmed that Big Walk falls on the wrong side of that line. If your concern is specifically whether the game used generated art, writing, or other creative content, the developer’s public statement says it did not. If your concern is disclosure, the studio says Steam did not ask for code or chatbot usage and that the public statement was made because the issue had become a matter of public interest.
The larger unresolved question is whether indie studios can maintain player trust with case-by-case explanations, or whether they will need clear AI policies before development begins. House House has already called the absence of its own policy a failure. For a scene built on craft, transparency, and the thrill of finding small teams doing surprising work, that may end up being the most durable lesson from the Big Walk developer AI controversy.