100% of Japanese Online Game Developers Now Use Generative AI, JOGA 2026 Report Finds

What the Report Actually Covers

Before drawing broad conclusions, the scope matters. JOGA defines “online games” narrowly: domestic Japanese titles played via the internet, regardless of device. Console games, PC games, and offline mobile games are excluded. The 100% adoption figure applies only within that defined segment.

That boundary is worth keeping in mind. It means the data reflects a specific, internet-connected slice of Japan’s game development landscape — not the industry as a whole.

Which Tools Are Leading the Stack

Among the generative AI tools in use, three names dominate:

• Google Gemini — used by 94% of surveyed companies
• Anthropic Claude — used by 84%
• GitHub Copilot — used by 76%

The spread across these three tools is notable. Gemini leads by a significant margin, but Claude’s strong second position suggests that developers are not simply defaulting to a single provider. GitHub Copilot’s presence confirms that code-generation workflows are embedded alongside content and analysis tasks.

The mix of Gemini, Claude, and Copilot also highlights how different tool types can coexist in development stacks.

What Developers Are Actually Using AI For

When asked which tasks they delegate to generative AI, the most common responses were analysing user preferences and predicting user behaviour. These are data-heavy, iterative tasks — exactly the kind of work where AI tools can process volume at a speed no human team can match.

This framing is useful. It positions generative AI in this context less as a creative replacement and more as an analytical layer sitting beneath game design and live operations decisions.

What Players Think

The report also surveyed players, not just developers. Two concerns surfaced repeatedly:

1. That games infringing on copyright are more likely to appear as AI use increases.
2. That future games will converge toward similarity — that AI-assisted development will produce a homogenised output.

These are not fringe concerns. They reflect a real tension between the efficiency gains AI offers developers and the creative diversity players expect from the medium. Whether those concerns prove accurate will depend heavily on how studios choose to use these tools, not simply whether they use them.

These concerns also fit into a wider pattern of public AI backlash as adoption expands.

The Disclosure Gap

The JOGA findings align with a broader pattern emerging across the global games industry. Google Cloud’s global director for games Jack Buser stated earlier this year that roughly nine in ten game developers globally are already using AI — but that many studios are reluctant to say so publicly, given how divisive the topic remains with players and press.

His observation points to a meaningful gap between actual adoption and disclosed adoption. Surveys that report lower figures — around 40–50% — may be measuring willingness to acknowledge AI use as much as they are measuring use itself.

The JOGA report, focused on a defined domestic segment with a structured annual methodology, may simply be capturing a more honest picture than broader, less structured surveys tend to produce.

What This Means for Tool Selection

For anyone evaluating AI tools for game development workflows, the JOGA data offers a few practical signals:

• Gemini, Claude, and Copilot are not just popular in general — they are the tools a mature, production-focused industry has converged on for specific reasons. Understanding what those reasons are is more useful than following the headline adoption number.
• Analytical and behavioural tasks appear to be the primary entry point for AI in game development, not generative content creation. That distinction matters when assessing which tools to prioritise.
• Player perception is a live variable. Studios that use AI without a clear communication strategy are accumulating a trust deficit that may become harder to manage as awareness grows.

The 100% figure is a useful signal. What it signals most clearly is that the question for Japanese online game developers is no longer whether to use generative AI — it is which tools to use, for which tasks, and how to be transparent about it.