House House Admits Using Generative AI to Fix Big Walk Technical Bugs
The developer confirms limited LLM usage for code troubleshooting following pressure from IGF jurors.
1Limited AI Implementation Confirmed
House House, the studio behind Big Walk, officially confirmed using Large Language Models during development. They stated this technology was used sparingly to resolve specific technical bugs. The studio emphasized that well under 0.1% of the game's total code involved LLM assistance, strictly limiting its application to solving complex programming challenges.
2Solving Rendering Pipeline Issues
The decision to use an LLM stemmed from persistent issues with Unity's rendering pipeline. Specifically, the team struggled to cull unnecessary shadow casters behind the player, causing performance drops at sunrise and sunset. After a Unity engineer deemed a fix impossible, the team turned to an LLM, which provided a ten-line code solution.
3Pressure from IGF Jurors
The disclosure followed a statement from 19 Independent Games Festival jurors who pledged to decline voting for any game utilizing generative AI. While the jurors did not name the title, discussions regarding Big Walk's eligibility surfaced during the judging process. House House subsequently updated their FAQ page to address the growing public interest.
4Mandatory Disclosure Requirements
House House noted that they had not previously disclosed this usage on Steam because the platform does not require it. However, disclosure became mandatory for their IGF submission. Bruno Diaz, a Pathologic 2 writer and juror, confirmed that judges were made aware of the LLM usage specifically because of these strict IGF rules.
5Voluntary Choice and Policy Failure
The studio admitted that the decision to use AI was voluntary rather than a management mandate. They acknowledged that Big Walk was developed without an internal AI usage policy, labeling this absence a failure. Moving forward, the developers committed to establishing a more comprehensive policy regarding the use of such emerging technologies.
6Stance Against Normalization
Despite using the tools, House House clarified that their actions do not constitute an endorsement of generative AI. They expressed concern that voluntary disclosures might contribute to the normalization of the technology. The studio apologized to those who feel the usage fails to meet the standards of industry professionals opposing AI-driven harm.
7Broader Industry Ethical Debate
The IGF jurors' statement highlights a growing movement among independent developers to push back against the perceived inevitability of AI. Signatories, including figures like Sam Barlow and Austin Walker, argue that AI-generated code is not exempt from the ethical harms associated with generative tools, urging a collective stand against industry-wide adoption.
8Future Implications for Developers
This incident underscores the tension between practical development needs and the ethical standards set by industry awards. As platforms like Steam remain neutral on disclosure, developers face increasing pressure from peer-led organizations to be transparent. The industry continues to debate whether minimal, functional use of AI tools is compatible with independent creative ideals.