Square Enix has announced a partnership with the University of Tokyo's Matsuo-Iwasawa Laboratory to automate 70% of its quality assurance and debugging tasks using generative AI by the end of 2027. This initiative aims to boost efficiency in game development and gain a competitive edge. The plan emerges from the company's medium-term business strategy amid broader AI adoption in the industry.
Announcement Details
Square Enix revealed its ambitious AI strategy in a progress report on its medium-term business plan, following the release of financial results. The Japanese publisher is focusing on generative AI to handle a significant portion of QA operations, with the goal of automating 70% of QA and debugging work by 2027. As stated in the presentation, "through the use of automation technology, [we] aim to improve the efficiency of QA operations and establish a competitive advantage in game development."
Partnership and Research
The project, titled "Joint Development of Game QA Automation Technology Using Generative AI," involves a research team of more than 10 members from both Square Enix and the Matsuo-Iwasawa Laboratory at the University of Tokyo. This collaboration originated from an internal AI-themed idea contest and aligns with Square Enix's push to promote AI utilization in Japan. The laboratory, dedicated to new technologies, seeks to foster innovation akin to Silicon Valley ecosystems.
Industry Context and Skepticism
Square Enix has been aggressive in AI application, having used tools like Azure OpenAI in development, including for Foamstars, though updates for that game ended earlier this year. The company has a history of exploring emerging tech, from NFTs and the metaverse to blockchain games like Symbogenesis. Recently, Square Enix joined other Japanese publishers in requesting OpenAI to stop training its Sora 2 model on their creative works.
However, opinions on AI in QA remain mixed. While some see potential in automating bug detection, veteran game artist Del Walker questioned its effectiveness, citing a bug in Marvel Rivals where a costume distorted a character's model. He asked on X, "What is GenAi really gonna say when something like this pops up? Can it even critically understand what it’s looking at to report the issue accurately?" QA's importance is underscored by past launch issues, like those in Fallout 76, highlighting the need for reliable testing despite labor-intensive processes.
