Japan’s Computer Entertainment Supplier’s Association (CESA) released a preview of this year’s CESA Video Game Industry Report at Tokyo Game Show 2026 (which the organization runs). Among various insights into the domestic game industry, the report found that 85.8% of Japanese game developers now use generative AI in their work. Among them, 63% of developers said they use it on a daily basis, while the remaining 22.8% use it occasionally (via Denfaminicogamer).
The report is based on a survey that targeted game developers and CESA member companies (the list includes a large number of prominent developers like Capcom, Sega, Level-5, and Konami). While the preview did not disclose the most common uses of generative AI, the most commonly cited benefits developers expected were “improved operational efficiency and productivity”, followed by “shorter development cycles” and “reduced development and operational costs.”
When asked about what measures they had in place for using generative AI, the most common response among developers was “conducting human verification, correction, and supervision.” The majority of responses cited practices such as specifying or restricting the tools used and avoiding the direct use of generated output.
Generative AI usage in game development has spiked from last year’s CESA Video Game Industry Report, which was estimated at 51% among CESA member companies and game devs. In the 2025 report, the most commonly cited use of AI was the generation of visual assets and images, followed by story and text generation, and finally programming support. It was also reported that 32% of game companies enlisted the help of AI to develop in-house game engines.
Commenting on the new 2026 statistics, CESA executive director Tsutomu Masuda said, “This demonstrates that generative AI is being widely and steadily adopted in development environments.” He added, however, that concerns regarding infringement that arise from generative AI must be addressed by the game industry as it continues to use the technology. He noted that moving forward, a key challenge for the industry’s future will be how to enhance the quality of human-centered game development while balancing the incorporation of new technological possibilities with the protection of intellectual property rights.
I can believe it, but i also theorize that GenAI for Art will be more In-House over time due to potential issues with IP leaking out or copyright problems, and that Code Generation and Text will be very much higher than GenAI for Art due to it being less IP based.
Why would the go in-house when they got the American government who allowed Google to steal anything it wants for it’s LLMs? Nah, they’ll go full internet.
Because of the fact there is ramifications for GenAI being used in public vs being utilized in private. It’s like Cloud Storage essentially, there is things you would trust in public for Cloud Storage like Microsoft and others and they have your data, and hosting it on your own Private Company. You do know you can do GenAI trained internally correct? I mean what you are mentioning is already old news from almost a year ago already and things have changed.
I will kill myself👍
👍
Personally I see AI in code and tooling as something neutral or even positive. It’s on art and sound I think we should steer away from it.
I agree with you on that, AI in Coding and Programming is essentially just 1’s and 0’s with some text strings, so it really does not matter as much honestly. It’s a very monotonous thing. I think having a Hybrid Workflow like Developer and AI working together would be hugely beneficial as seen in recompilation and various other boilerplate tasks that would take weeks and instead take a few hours to prevent Crunch.
On sound I guess it depends on how it is made, like if it’s sort of like asking AI how to make a certain sound that is not easy to do otherwise or a sound that is impossible to make I get it. Art Generation has come a long way it’s great for prototyping designs and ideas. But for frontfacing tasks and marketing it looked very off as seen with Level-5’s Presentation, I think it needs a few years in the oven to be perfected.
And Capcom is on a generational run. Coincidence??
Huh? Past 10 years from capcom has been like remakes, remasters, and releases.
Besides maybe exoprimal? Flop.
Wilds? Absolute flop.
For this year, pragmata maybe? But thats one.
guess i’ll take up knitting or something
Abysmal.
shut up 👎
Other commenters are being anti tech so reactionary, no nuance acting like right wingers.
Ah, the Leftist, so up their own heads because they don’t understand nuance, assuming everyone is like them.
AI steals, not maliciously, because it’s not actually intelligent, it’s a curated search engine that collects all the data it can, without regard to actual ownership.
YES!!! get this to 100% asap!!!!!
It’s true. Most rpgm corn game use generative ai. Been playing those type of game in recent month.
Yup and the hypocrites like MoistCritikal, YongYea, Maximillian Dood, and others will shut up and consume theb gaslight when they are called out as always or run and hide. Expose these hypocrites who ruined the world.
AI steals, that’s the problem. It doesn’t create, it can’t. It’s not actually intelligent, it’s a search engine with access to the entirety of the internet where it can take whatever it wants, including copyrighted material.
But because you have no artistic skill, you don’t actually care.
This is one of the stupidest things I’ve ever seen, and is obviously written by someone who doesn’t understand how AI works, nor how people use it, nor how actual creators create.
Those who complain about AI today are too much like those that complained about 3D back in the day or Photoshop. I think what people dislike is low effort AI quality not truly AI in of itself. People need to make that distinction rather than saying AI is bad or AI is awful. Because it varies.
No… Its not. Not even remotely close with your 3D comparison. Which means you don’t even understand the gaming industry or public opinion. 3D was just a visual change and took a lot of resource so only triple A could afford to do.
In the last few years and currently, people don’t like genAI because of the resources it takes up, the impact physically, digitally, and emotionally. Trained on everything that was datascraped in the past 20 yeara online, from art, personal profiles, any form of writing, blogs. Its all stolen in an amalgamation to copy and reproduce things. When people found this out more publicly, that is the issue.
It cannot do things outside whats shoved in it. It doesn’t “learn” and even thinking models isnt “thinking”, its just a checklist. Every genAI can only do what its built to do, which is why only this year, theyve connected it to other program to do things intermittently like 3D model and use a photo editor. Newest chatgpt6 can essentially copy games and spit out crappier but very playable games now. Why couldnt you tell it to do it before in 4 or 5? Because they are not capable of actually learning from you. Its capacity of “mind” is no where near nor will it ever be in our life time, to make decisions like us or create like us.
Same thing was said even back then that Photoshop took away effort from actual Artists that took them years to do things that were “Hand drawn”, and also 3D was awful when it came out, it was extremely uncanny before people realized it was about those that basically used it correctly and put in effort or where extremely talented vs those that put in low effort work. Also let me just mention the whole “it scraps data from the Internet” is technically not true as it used to be if you look at the Supreme Court ruling of Bartz Vs Anthropic, in which yes Anthropic was caught pirating, stealing and scraping, but had to legally purchase works, pay artists, and had to do it in a normal manner. Now the industry is pivoting away from data scraping to Synthetic Data. Also I do not think you understand that Anthropic, OpenAI etc do not keep images stored on a data base, but basically learn how words, pictures, 3D is done and train it from that. If this were the case every single Video Generator, Image Generator, and even text would be the exact same result every single time. I understand your outrage, and your emotional anger, but try to do research please.
You have no idea what the difference do you?
AI steals, from everything, because it often has access to the internet. Now, I know because you don’t create, just consume, that this means nothing to you. But to actual artists who want to create something unique and maybe sell it to others, will find that every big box company will have stolen all the assets and will sell that as their own, without needing to even credit where they got it. It’s theft.
Unlike Photoshop or 3D. But you’re not going to believe or even care about this. You just want dump on people you hate.
” Now, I know because you don’t create, just consume, that this means nothing to you. ” Way to assume you know anything about me and resort to ad hominem attacks and resort to emotionally charged arguments. Calling everything theft without really understanding how GenAI actually works is just basically making your side look bad as a whole.