Stellar Blade's director says AI can see 'one person can perform the work of 100 people' but that won't wind up in job losses, somehow

We're not even a few steps deep into January yet, and the AI crossfire of 2025 continues. I say crossfire because it's unclear where any of this is going to land. AI's as big as the shift to 3D, says Ubisoft—smack-bang in the middle of an AI bubble. Half of Japanese game developers are using it! Also, Arc Raiders developer Embark is still lukewarm on the stuff despite controversially using it during its development cycle.

Larian, ostensibly one of the more influential studios out there, got thrown in deep water for AI use during concept art iteration and then decided to not do that, despite continuing to tinker in other areas. How useful this tech actually is on a practical level (you know, when software companies aren't mandating employees use it or complaining that nobody likes it) remains to be seen.

Stellar Blade director Hyung-tae Kim reckons he knows the future, though, speaking at the 2026 Economic Growth Strategy briefing on January 9 (thanks, Automaton). The briefing, an official government shindig overseen by South Korea's president, invited Kim to share his thoughts on the future of AI in the gaming industry.

The problem, as Kim puts it, are Chinese developers—who are a direct competitor, seeing as the studio owes 80% of its revenue to overseas markets (the following quotes are machine-translated): "We devote around 150 people to a single game, but China puts in between 1,000 to 2,000. We lack the capacity to compete, both in terms of quality and volume of content."

Kim then, in what I can only describe as a tremendous feat of mental acrobatics, states that AI adoption will help the South Korean games industry compete with these big stonking overseas companies, so that "one person can perform the work of 100 people", and yet somehow this won't lose people their jobs. The implication being that South Korean studios are already thin on staff anyway, so AI will be more of an equalizer than a job-destroyer.

It's true that in some cases, AI might allow one person to be a decent peg more productive—it'll replace some jobs entirely—but the claim that AI adoption can supe up devs across a company by a factor of 10 to one just isn't substantiated. Especially if you're painting in broad strokes.

If anything, the last year of AI humiliation makes me feel like the gloss has fallen off the technology. Sure, it'll have its uses, in the same way procgen does, and it might even speed up coding a bit (though "vibe coding" carries its own risks). But these use-cases are boring and dull, and not the kind of promises that get the South Korean government to invest in AI initiatives, as they are indeed planning to do in 2026.