Söderlund is careful to stress that AI is not a replacement for human creativity. “The beauty of video games, as with any art form, is that — at least so far — they can’t be built by an AI,” he says. “I hope they never can. The human element remains essential.” Instead, AI is being applied to the more mechanical, technical tasks that otherwise consume time better spent on design and craft.
He gives a concrete example from Embark’s experiments: rather than hand-animating or capturing motion for every frame, what if the team built a physical model, defined its attributes, and trained an AI to produce walking animations? In 2019 Embark published a video demonstrating that approach.
A post shared by Embark Studios (@embark.studios)
That experiment allowed characters to move “without any manual animation work at all.” Six years on, the tools have advanced considerably. Söderlund notes that the studio can now take a video — for example, a clip from YouTube — process it through their pipelines, and generate a usable 3D model from it.
He is candid that the output isn’t flawless: “That might sound like wizardry and I would be lying if I told you that it’s perfect — it isn’t — but it works, and that output is something we can then refine.” Embark initially aimed to produce content ten times faster, but Söderlund wanted to push the ambition far beyond that.
Söderlund recalls telling Embark co-founder Robert Runesson that he didn’t want to aim merely for tenfold speed-ups. “I know you’re going to think I’m a complete nutcase,” he said, “but I don’t want us to build things 10 times faster — I want us to build content a hundred times faster.” His point: setting a much higher goal forces you to rethink every assumption rather than make incremental tweaks.
So why has Arc Raiders experienced multiple delays? Söderlund points to an old default at big publishers: when schedules are tight, the immediate solution is often to “just put more people on it.” At EA, that could mean hundreds or even a thousand developers added to a project. Embark rejects that model as short-term and inefficient.
Instead, Söderlund argues the priority should be time to learn and iterate: testing systems, mining data, and understanding what the results mean. “It’s not about money or headcount,” he explains. “You need sufficient time to learn the things you need to learn. You can’t really rush that.”
Those remarks may not assuage critics who worry about AI’s role in creative work, but they do reveal the trade-offs Embark is weighing as it adopts new tools and adjusts its production approach.
Legendary Final Fantasy composer Nobuo Uematsu has said he has “never used AI and probably never will,” preferring the personal rewards of creating through his own struggles.
Source: gamesradar.com
