Australian developer House House has confirmed that generative AI was used "sparingly" during the development of its acclaimed multiplayer puzzle game, Big Walk. The admission comes in the wake of a controversy at the Independent Games Festival (IGF), where 19 jurors announced they would abstain from voting for any game that incorporated generative AI in its creation.

In a statement provided to Aftermath, House House clarified that large language models were employed primarily to assist with technical challenges and bug fixes during the game's final development stages. The studio estimates that "well under 0.1 percent" of Big Walk's codebase was influenced by this usage, and generative AI was not utilized in any other capacity. The primary application of LLMs involved researching issues with third-party code via chatbots.

The studio had not previously disclosed its use of generative AI, as they noted that platforms like Steam do not require such disclosures for code or chatbot usage. However, House House stated they are now making this information public because it has become a "matter of public interest."

The developers expressed concern that voluntary disclosures might contribute to the normalization of generative AI in game development, or its perception as an unavoidable tool. They voiced support for efforts to mitigate the negative impacts of generative AI on the gaming industry and the wider world, and apologized if Big Walk's minimal LLM usage did not meet the standards set by those concerned jurors. House House also acknowledged a failure in not having an internal AI usage policy during development and committed to establishing a more comprehensive policy moving forward.

This revelation could cast a shadow over Big Walk, a title that received a perfect five-star rating from Eurogamer in its review. The game is a celebrated indie title, and its use of generative AI has prompted significant discussion within the game development community.