The indie game world is arguing over a single chatbot-assisted bug fix. House House, the studio behind the co-operative puzzle exploration game Big Walk, has publicly disclosed that it used large language models, or LLMs, sparingly late in development. The disclosure came in the middle of a broader dispute over AI at the Independent Games Festival Awards, and it has pushed the issue into the open.
Why are some IGF jurors refusing to vote for games with generative AI?
The fight began on Monday, when 19 members of the IGF's jury of more than 50 people released a signed statement. The jurors said they were “declining to vote for any game that uses generative AI in any capacity.” They pointed to what they called “a significant increase in the number of submitted games developed with generative AI tools,” and listed the technology's ecological, economic, psychological and labor harms as reasons for their stance. They also asked the IGF to adopt “a strong anti-AI policy for future years that disallows games containing any AI-generated work.”
What AI did the Big Walk developer use?
The statement did not name a game. By Tuesday morning, though, Big Walk had become the center of the debate. In a statement given to Aftermath, which was also added to the game's FAQ page, House House said that “LLMs were used sparingly” near the end of development “to help work through some difficult technical challenges and bugs.” The studio said that use accounts for “well under 0.1% of Big Walk's code.”
The problem started with a Unity engine bug that slowed the game down by generating “infinite length” shadows in certain situations. A Unity engineer determined a fix “wasn't possible,” according to House House, so the team consulted an LLM chatbot. The chatbot suggested a possible fix, which the developers “had to modify… extensively” into the 10 lines of code that shipped in the final game. At least one social media user has said the bug was fixable with significant human effort, though.
Why did House House disclose its AI use now?
House House said it did not disclose this usage when Big Walk launched on Steam earlier this year, because Valve “does not request disclosure of code or chatbot usage.” Valve's AI disclosure form focuses on AI-generated content and does not cover what it calls “efficiency gains” from “modern game development environments [that] have AI powered tools built into them.” The IGF makes no such distinction. That is why the studio mentioned the AI-assisted fix when it submitted the game for awards consideration. The studio says it is making that disclosure public now because “it has become a matter of public interest.”
“This is not an endorsement of, or argument for, the use of AI tools, however minimal,” House House wrote. The studio added that it worried voluntary disclosures “might contribute to the normalization of this technology, or its treatment as ‘inevitable.’” It also apologized that Big Walk's LLM use “may not meet their standards.”
Some IGF judges had felt a close connection to the game, which made the news harder to take. In a blog post, jurist Austin Walker called Big Walk “an achievement” that was “remarkable” and “beautiful,” and praised puzzles that “capture a range of play that is largely abandoned by the digital form.” Fellow jurist Damon Reece wrote that the game is “seriously special to me.” Even so, Walker and Reece were among those who said that even minimal AI use outweighed their respect for the work itself. Reece wrote, “We have our hands on the lever here,” and called staying silent “rank cowardice.”
This is not the first time AI has affected awards judging in games. Last year, the Indie Game Awards, which is separate from the IGF Awards, rescinded honors for Clair Obscur: Expedition 33 after its developers admitted that some in-game textures were made with AI tools. Those textures were later patched out. At the time, the Indie Game Awards said it has “a hard stance on the use of gen AI throughout the nomination process and during the ceremony itself.”
Fans have also reacted sharply. Big Walk's Steam page has been flooded with more than 400 negative reviews over the past few days, though Steam's tools against review bombing appear to be limiting the damage to the game's overall “Very Positive” rating.
Others think the reaction has gone too far. IGF judge Liz Ryerson wrote on social media that the AI use was not personally disqualifying for her. “The fact that many acted like there aren't any levels to this and it's all the same is bothersome to me,” she wrote. Dominic Tarason, a Steam Hype staff writer and prolific Steam curator, argued that the studio should not be “memory-holed.” He called Big Walk “very mildly compromised art” that is still “absolutely, unquestionably art and very human.” “They asked a chatbot for two minutes how to fix a bug,” he wrote. “That's a ‘That's lame, don't do that again’ failing.”
Critics of the studio's choice see a different problem. AI researcher Mike Cook wrote on his blog that House House is not laying off developers to replace them with AI, but that “we can also clearly see how their actions are part of the same problem.” Game developer, author and professor Brendan Keogh took a middle view. He wrote that an “extreme zero tolerance” approach is “ultimately counterproductive,” and warned that harsh reactions could discourage developers from being open about their AI use.
As AI tools become more common in building major game engines, avoiding every game touched by LLM-generated code may get harder, especially if the use goes undisclosed. Still, some people see a firm public line as worth holding. Narrative designer and writer Hannah Nicklin wrote on social media, “I think zero tolerance is a viable political position.”
Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.