“Our guiding principle when it comes to using AI for game development is that it needs to assist the creator, not the creation,” Jacquier adds. With a “human-in-the-loop” approach, Jacquier says that AI will not replace or compete with developers, but will facilitate and optimize some aspects of their work, or open new possibilities of creation for them.
“What has constantly changed over the past years is both the maturity and increasingly easy access to more advanced forms of applied AI, and consequently the number of applications for AI in our workflows, such as generative AI,” Jacquier explains. Still, he says that there are far-reaching challenges, such as creating a common framework to train AI models for video games.
“While designing a game, if you tell an algorithm that an island can only be next to water or to other islands, it can get that set of rules and continue creating forever,” Movilla explains. As an example, he describes huge automatically and procedurally generated maps in video games that are already created by tools that many of us would call AI.
When developers rely on AI in various stages of their game development, those algorithms become “smarter.” They collect and propose ways to deal with new data, and they keep track of how players react to certain game features. The latter is usually measured by what’s called game telemetry—the data that is sent back to developers when you play their games.
“Every single behavior that you're doing, when you're playing these games, is being monitored,” says Movilla. He explains that the data obtained by telemetry offers developers opportunities to change the game to improve the player experience and even make more money, especially in the case of live-service games or games with microtransactions.