Over Half of Japanese Game Studios Now Use Generative AI, Says CESA Report

Japan’s Computer Entertainment Supplier’s Association (CESA), the group behind Tokyo Game Show, has released a preview of its upcoming 2025 CESA Video Game Industry Report. The findings, shared on September 25, highlight just how deeply artificial intelligence has begun to shape game development in Japan.

How Widespread Is AI in Japanese Game Development?

According to the report, which surveyed 198 Japanese game companies between June and July 2025, 51% of studios are already using generative AI in some form.

The most common application is the creation of visual assets — things like characters, background art, and props. Story and text generation came next, followed by programming support.

Interestingly, about 32% of companies are even using AI to help build their own in-house game engines, showing that this isn’t just about surface-level art.

The survey included a wide range of CESA member companies, from industry giants like Capcom and Sega to smaller and mid-sized developers. That mix gives us a pretty balanced snapshot of how the Japanese industry is approaching AI.

Some studios have been transparent about their AI experiments. In early 2025, Capcom’s technical director Kazuki Abe explained in an interview with Google Cloud Japan that the studio has been testing Google’s Gemini Pro, Gemini Flash, and Imagen models. These tools are mainly used for brainstorming and prototyping background assets — think TVs, furniture, or filler objects that make environments feel alive.

Sega has taken things a step further by forming an internal Generative AI Committee. This group is tasked with systematically integrating AI into development workflows, covering areas like image generation, motion capture, and even code.