Unity, the cross-platform game engine company that provides a range of tools and services for creators to produce games, plans to integrate generative AI into its engine.
The company’s senior vice president of artificial intelligence, Danny Lange, tells Fast Company that the real-time game development platform is embracing AI to assist game developers that utilize Unity’s engine to produce higher-quality content quicker.
“We look at things like AI as very exciting ways to imagine how you open up the opportunity for professional creators to be even more productive,” Lange says.
Game development has long been known as a high-stress, high-effort world (which in part explains recent drives for unionization across the industry). But AI also offers new tools to broaden access to those wanting to create their own games. “[It allows] a whole new wave of people who might have some ideas to have access to a set of technologies that would enable them to be very successful,” Lange says.
AI has been part of Unity’s tool kits for a while: The company works with Ziva Dynamics, another firm, on its face trainer, which uses AI to produce high-quality face animations for characters. With the face trainer, something that had once taken artists three or four months to complete, can now be done in minutes.
Unity also already operates Barracuda, a neural network-inference tool that can identify individuals and objects within footage by using image recognition, as well as AI agents that can test games developed within the engine to check for issues—an automated version of quality assurance.
Lange declined to share how much total investment Unity has made in the AI space, beyond saying it was a “pretty reasonable” amount of cash. “We have a very significant number of engineers, product managers, and designers working on AI tools,” he says.
In order to capitalize on that movement, Unity will be ensuring AI features can be integrated within the Unity Editor, an app within Unity that allows users to edit items. “You’ll be able to go to our assets store, just like you have for every other category of thing that gets added into games, and use them,” he says. “In addition to the work that we’re doing to develop our own tools, we’re going to make it very easy for this huge wave of innovation and tools to be available inside of our asset store in our marketplace.”
The end goal is to speed up the ability to create high-quality games. Developing games takes time, says Lange, and “the vast majority of that time is filled with creating the content, and the content is a massive limiter.” By being able to automate or outsource at least some of that work to AI, the objective is to improve efficiencies in producing games.
Lange also hopes that, in the long run, AI can be deployed not just during a game’s development, but also on players’ devices, to reduce the burden on cloud-delivery networks for games. Rather than games pinging centralized servers to trigger AI that can change the game, those decisions could be made on the fly on user devices. That would mean non-playable characters’ decisions, or the environments in which the game world exists, would change dynamically based on generative AI processed on a player’s device. “We are trying to create the technology where game designers could actually think about variation and AI as part of the game loop itself, that it gets to players, and make that work in a way that’s both economical and fun.”
Nor does Lange believe Unity will be alone in integrating AI into its products. “I think pretty much every creator tool is going to have an AI underpinning it over some time frame,” he says. “I think every software developer is going to be using some AI tools to sort of help develop it, I think every graphics artist is going to be using it, and every writer is going to be using it.”