News
All Japanese Online Game Developers Now Use Generative AI, JOGA Report Finds

A new Japan Online Game Association report finds that 100 percent of surveyed Japanese online game companies now use generative AI, but mostly for analyzing player behavior rather than generating art or story content.
Contents
The Japan Online Game Association, together with Kadokawa ASCII Research Laboratories, published its annual online game market report on July 10, and the findings show that every single surveyed company reported using generative artificial intelligence in its work. At first glance that sounds like a record-breaking expansion of AI across the entertainment industry, but the report's details point to something different from the usual fears about replacing artists or writers.
The survey covers only companies that belong to JOGA, meaning firms that make online games in the sense of titles accessible over the internet regardless of device. It doesn't cover the entire Japanese games industry, and it excludes offline console and PC productions as well as standalone mobile games without a networked component. Even with that narrower scope, a 100 percent result is striking, since similar surveys by CESA, Japan's computer entertainment supplier association, put AI adoption at roughly half of companies not long ago.
What the AI is actually used for
Headlines in the Western press sound sensational, since the intuitive assumption is that 100 percent generative AI adoption in games means a flood of art and text produced by models. JOGA's report paints a different picture. Companies most often reach for AI to analyze player preferences and predict their behavior, analytical work that feeds directly into design decisions and gameplay mechanics.
Other uses cited in the survey include coding assistance, drafting project documentation, and supporting select stages of production. Generating art or scripts, the use that worries players most, appears in responses less often than these analytical tasks.
Gemini and Claude beat ChatGPT
The breakdown of tools used by Japanese developers confirms a trend already visible elsewhere in the region: Google Gemini, cited by 94 percent of respondents, is the most popular model, not OpenAI's products. Anthropic Claude follows closely at 84 percent, with Microsoft's GitHub Copilot rounding out the top three at 76 percent among surveyed companies.
These figures line up with earlier signals from the Japanese games market, where developers of networked titles have gradually moved away from ChatGPT in favor of rival models from Google and Anthropic. Gemini and Claude's dominance among developers may stem from tighter integration with development tools and both companies' aggressive push into Asian markets in recent months.
Player concerns
The second part of the JOGA report looked at the perspective of players themselves, not just the companies making the games. Respondents most often flagged two concerns tied to AI's growing role in production: the risk of copyright infringement in machine-generated content, and games becoming increasingly similar to one another if developers lean too heavily on the same tools and models.
The worry about product uniformity isn't unfounded. Since nearly every company relies on the same handful of models to analyze player preferences, mechanics and reward systems optimized around the same behavioral patterns could over time make titles from different studios look and feel alike, even if none of the games contain content generated directly by AI.
What it means for the industry
For Polish studios and publishers working on online games, the JOGA report signals that the most valuable use of generative AI in this segment may not be asset creation at all, but player-data analysis at a scale that previously required large analytics teams. That also shifts the questions investors and regulators should be asking the industry: not just where the art and text come from, but how AI models are shaping the design of reward mechanics and player retention.
The finding coincides with a broader debate over labeling AI-generated content in games, one also playing out in Europe, where regulations are starting to require transparency toward players. Japan's market, with formally 100 percent AI adoption among JOGA member companies, could become a reference point in that debate, though the report itself notes it covers only a narrow segment of online games, not the Japanese industry as a whole.
