Resident Evil and Street Fighter series developer Capcom is experimenting with introducing new technology, including generative AI, to tackle the ballooning costs and man-hours required for game development. In a recent interview with Google Cloud Japan, Kazuki Abe, a technical director at Capcom, gave some specific examples of what this involves. Based on his explanation, it doesn’t seem like Capcom is trying to use AI to generate anything directly related to gameplay, stories or character designs.
According to Abe, one of the most time-consuming and labor-intensive parts of game development is coming up with the “hundreds of thousands of unique ideas” needed to create the in-game environment. For example, if you want to put a TV inside of your game, you can’t just use an existing product as is – you need to think of a fictional TV design from scratch, including the manufacturer’s logo and everything else about the object.
You must log in or register to comment.
For example, if you want to put a TV inside of your game, you can’t just use an existing product as is – you need to think of a fictional TV design from scratch, including the manufacturer’s logo and everything else about the object.
This is such a bullshit problem. No other media has this issue. This wasn’t a problem before and never should’ve been allowed to become a problem. Real objects have populated the scenes of movies for decades, artists have painted scenes rendering real clothes and weapons and carriages and all with accurate shapes for centuries, before the 2000’s video games seemed to have no issue doing the same. Something changed and now we’re left with absurd assets.
Rendering real objects without any limitations or requirement for legal agreements should be a right of any artist in any field.
That’s sadly not how the capitalist implementations of fair use and copyright in general work.
It needs a rework. It’s not in the people’s interest to have these limitations. The US shouldn’t be leading the copyright topic any longer, maybe the EU could fix this.
According to Abe, one of the most time-consuming and labor-intensive parts of game development is coming up with the “hundreds of thousands of unique ideas” needed to create the in-game environment. For example, if you want to put a TV inside of your game, you can’t just use an existing product as is – you need to think of a fictional TV design from scratch, including the manufacturer’s logo and everything else about the object.
I can just go into the asset browser and find a whole slew of TVs I can just add to my game and either use them as-is, or tweak the look a bit to fit my aesthetic choice and not need to reinvent the wheel every damn time. It’s such a small thing that’s just filling out the background, that adding details like a fictional logo aren’t necessary at all, unless it’s actually relevant to the plot somehow. What the fuck is Abe smoking with that example?
It does make sense if you replace the word time with money. Paying all those artists to make all those little details for every little thing is costly. But AI will do it for free.
On the subject of existing asset use… I played yonder: the cloud catcher chronicles, and I am fairly certain the sound for rain in that game is actually the sound of something gently sizzling in a pan.
It’s super close, but -just- wrong enough for me to notice. And then be mildly bothered by it for the next 40 hours of playtime.
People will call you an asset flip if you do that. Plus I don’t know if all of these games use engines compatible with “the” asset browser
“The” asset browser could be any number of sources, some of which are not coupled to any specific engine. AAA games do also use already made shit, especially when it’s an inconsequential background item like a TV. They just don’t use them for the entire game.
Aren’t ideas like, the easiest part of making a game??
I thought this was The Onion and now I’m disappointed.
Pretty sure the actually developers at Capcom have more creative ideas than the AI they would use
Cool, cool. More jobs to disappear.
They could poll fans for free. They’d find out what their market craves and build hype st the same time.
Alas. They’ve decided to champion AI drivel. Which virtually guarantees fewer sales and fan backlash.
I got an idea for a fictional TV. It’s a black rectangle with a moving picture in the middle. There’s a logo on it that almost says the name of a real TV brand but in a slightly different typeface than they use and one or two of the letters is changed.
This is a revolutionary idea that nobody has ever had before, which if implemented will actually negate the need to use AI to create fictional TVs for us.