House House, the developer behind the cooperative puzzler Big Walk, has revealed that generative AI tools, specifically large language models (LLMs), were used "sparingly" during the final stages of the game's development to address technical challenges and bugs. This disclosure follows a statement from jurors at the Independent Games Festival highlighting a significant increase in AI-assisted game submissions and sparking debate around the use of such tools and the ethical implications of training data.

In an update to its website's FAQ, House House clarified that in some instances, snippets from LLM responses were used as a starting point for game code, modified by developers. The studio estimates that "well under 0.1% of Big Walk's code" involved this usage, and that generative AI was not employed in any other capacity. The developer provided an example where approximately ten lines of code, generated by an LLM, were used to resolve an issue that proved difficult to fix manually or with direct assistance from Unity engineers.

House House stated that it did not disclose AI usage on Steam because the platform does not require it, but opted to do so now as it has become a "matter of public interest." The studio also acknowledged that developing Big Walk without an internal AI usage policy in place was a "failure" and committed to establishing a more comprehensive policy for future projects.

The developer emphasized that this statement is not an endorsement of AI tool usage, regardless of its minimal implementation. Big Walk has also gained attention for its developers' decision to withhold the game from Xbox consoles in support of the pro-Palestinian Boycott, Divestment, and Sanctions movement.