100% of Japanese online game developers use AI, according to survey
A new industry report claims every single Japanese online game company surveyed uses generative AI. But the number comes with big caveats, and what they’re using it for is far more boring than you’d think. Meanwhile, a Google exec says the real story is how many studios won’t admit it.
A new industry report has produced one of those numbers that stops you cold: 100% of Japanese online game companies surveyed say they use generative AI tools. Not most. Not a growing majority. All of them.
It’s a striking stat, and it’s also one that deserves a closer look, because both the fine print and the actual use cases tell a more complicated story than the headline number suggests. Here’s what the survey really found.
What the report says
The figure comes from the JOGA Online Game Market Research Report 2026, published July 10 by the Japan Online Game Association and Kadokawa ASCII Laboratories. JOGA has run this survey annually since 2004, making this its 22nd edition, and the 2026 report added questions about how developers and players feel about AI.
According to a preview published by Famitsu, every company that responded reported using generative AI in some capacity. The most widely used tool was Google’s Gemini at 94%, followed by Anthropic’s Claude at 84% and GitHub Copilot at 76%. Those aren’t small numbers either, they suggest most companies are using several tools at once rather than standardizing on one.
The caveats matter a lot
Before anyone declares Japanese game development fully automated, the fine print deserves attention. JOGA stresses that this report covers only domestic Japanese games “played via the internet regardless of device.” That explicitly excludes console games, PC games, and standalone offline mobile games.
So this isn’t a survey of Capcom, FromSoftware, and Nintendo. It’s a survey of Japan’s online-game sector, a real and substantial industry, but a specific slice of it. The other thing worth noting: the reporting on the preview doesn’t disclose how many companies actually responded. “100% of respondents” is a very different claim if that’s 15 companies versus 150, and without a sample size, the number is more of a directional signal than a hard fact. It’s genuinely interesting. It just isn’t the sweeping verdict the headline implies.
What they’re actually using it for is refreshingly dull
Here’s the detail that deflates a lot of the panic. When asked what tasks companies were most eager to hand to generative AI, the top answers weren’t “drawing our characters” or “writing our scripts.” They were “user preference analysis” and “user behavior prediction.”
In other words: figuring out what players like and guessing what they’ll do next. That’s data analysis, the deeply unglamorous back-office work of running a live online game, and it’s the sort of thing online-game companies have been doing with software for well over a decade. The new part is that they’re pointing generative AI at it. So while “100% of developers use AI” sounds like an industry-wide creative takeover, the leading use case in this particular survey is closer to a very expensive spreadsheet. That’s not nothing, but it’s not what most people picture.
The exec who says everyone’s doing it and hiding it
The more provocative claim comes from elsewhere. Jack Buser, Google Cloud‘s global director for games, told Mobilegamer.biz in April that practically every major game studio is already using AI, they’re just not saying so out loud.
“I think what players don’t realise is that their favourite games right now were already built with AI,” Buser said. “Those games have shipped.” He argued the wildly varying survey numbers across the industry don’t measure adoption at all, they measure honesty. “You’ll see other surveys from other organisations that have that more around like 40-50%,” he said. “You might ask yourself, well, that’s still a large number... What’s that gap? And that gap is basically the developers’ willingness to tell you.” His pitch for Google’s own tools, like Gemini, is that they remove “the drudgery and repetitive, low value work.”
Take that with a grain of salt, and a data point
Buser’s claim deserves scrutiny for an obvious reason: he works for Google, Google sells AI tools to studios, and this very survey found Gemini sitting at the top of the adoption chart at 94%. An executive whose job depends on AI adoption telling you AI adoption is universal is not a neutral observer. That’s not a knock on him, it’s just context worth holding.
That said, there’s real evidence backing him up. A recent survey of Japanese creative professionals found 59% of companies used AI, and of those, 71.4% did not actively disclose it. That’s a remarkable number, and it supports Buser’s core point precisely: the gap between AI use and admitted AI use is enormous.
The Tomb Raider AI-voiceover backlash and similar incidents explain why studios stay quiet. Whether that silence is prudent or dishonest is the argument the industry hasn’t had yet.
The players, meanwhile, have concerns
The same JOGA report asked gamers what worried them about AI, and their answers are worth sitting with. The two most frequently cited concerns were copyright infringement and the possibility of “all games starting to look alike.”
That second one is the interesting one, because it’s not a legal or ethical objection at all. It’s an aesthetic one. Players aren’t just worried about where the training data came from, they’re worried about ending up in a world where everything is competently generated and nothing surprises them. And notably, this concern is coming from the audience of an industry that, per this same survey, has already gone all-in.
The wider picture is messier than 100%
Zoom out and the numbers scatter wildly, which tells you something. A CESA survey last September, covering 54 companies including Capcom, Konami, FromSoftware, Square Enix, and Sega, found just 51% using AI, mostly for visual assets, text generation, and programming support, with 32% using it to help build in-house engines. A Harris Poll of 600-plus studios across the US, South Korea, and Europe put the figure at 87%.
Individual positions vary just as much. Level-5 CEO Akihiro Hino has said roughly 80-90% of the studio’s code is now written by AI. Nintendo has taken a famously hard line against it. Xbox recently ditched a Copilot feature, with its CEO saying console players “aren’t excited about that.” So the honest summary isn’t “the industry has decided.” It’s that adoption is real, uneven, concentrated in unglamorous tasks, and wildly inconsistent in how openly anyone talks about it.
Japan’s 100% AI number: what it comes down to
The 100% figure is a genuine data point and a lousy headline. Read carefully, it says that in one specific corner of Japanese gaming, generative AI has become table stakes, mostly for analytics rather than art, on a sample nobody’s disclosed. That’s meaningful. It’s not the robot uprising.
The more durable insight is the one Buser stumbled into, even if he had a product to sell while saying it: the numbers everyone’s arguing about may be measuring disclosure more than deployment. Studios are using these tools, players have real concerns about copyright and sameness, and the industry has largely decided that the safest move is to say nothing. That’s not a sustainable arrangement. Eventually somebody has to tell the audience what’s actually in the box.
The scariest thing in this survey isn’t that everyone’s using AI. It’s that most of them would rather you didn’t know.
Want More Clownfish TV?
This article was brought to you in part by The Reefers of more.clownfishtv.com. Free subscribers get articles like this one in their inbox. Paid subscribers get the full Clownfish TV podcast feed, livestreams, and members-only episodes that never hit YouTube.
D/REZZED is part of Clownfish TV. For more news, views, and rants on gaming, tech, and pop culture, watch @ClownfishTV on YouTube and find the podcast on Apple Podcasts, Spotify, and iHeart.
Article compiled and edited by Derek Gibbs (entertainment editor) and the Clownfish TV newsroom.
Hat Tips:
Automaton West and VGC (July 2026), verified for the survey itself (the JOGA Online Game Market Research Report 2026 published July 10 by the Japan Online Game Association and Kadokawa ASCII Laboratories, the 22nd annual edition running since 2004, Famitsu’s preview reporting a 100% generative-AI adoption rate among surveyed companies, Gemini at 94%, Claude at 84%, and GitHub Copilot at 76%, the top delegated tasks being “user preference analysis” and “user behavior prediction,” JOGA’s explicit scope limit to domestic games played via the internet excluding console, PC, and offline mobile titles, and players’ top concerns being copyright infringement and “all games starting to look alike”)
Mobilegamer.biz (via VGC) (April 2026), verified for the disclosure claim (Google Cloud global director for games Jack Buser stating that practically every major studio uses AI in development but not all are comfortable disclosing it, his “their favourite games right now were already built with AI” and “those games have shipped” comments, his argument that the gap between 40-50% survey results and actual usage reflects “developers’ willingness to tell you,” and his description of Google’s tools removing “the drudgery and repetitive, low value work”)
Automaton West, PC Gamer, GameSpot, and Game Rant (2025-2026), verified for the comparative context (the CESA 2025 Video Game Industry Report surveying 54 companies including Capcom, Konami, FromSoftware, Square Enix, and Sega and finding 51% using AI with 32% applying it to in-house engines, a Harris Poll of over 600 studios across the US, South Korea, and Europe finding 87% adoption, a survey of Japanese creative professionals finding 59% of companies used AI with 71.4% not actively disclosing it, Level-5 CEO Akihiro Hino’s statement that roughly 80-90% of the studio’s code is AI-written, and Nintendo’s opposition to AI use)