Warhorse Studios fired the translator for Kingdom Come: Deliverance 2 and announced that AI will replace all human translation work going forward. The translator went public. This is not speculation about the future. This is happening right now.
According to Kotaku, this is one of the first major game releases to completely replace professional translators with large language models. Translation was supposed to be one of the knowledge work jobs that required human expertise. Turns out companies don't care.
What Actually Happened
The translator had been working on the game's localization. Then Warhorse Studios decided to pivot to AI-based translation for all languages. The translator was let go, and the studio confirmed that LLMs would handle translation work moving forward.
This is a major AAA game release with a substantial budget. The choice to use AI translation wasn't about a lack of resources. It was about cost optimization and speed.
Game translation is technically demanding. It's not just swapping words between languages. It requires understanding context, cultural references, humor, and maintaining character voice across different languages. It's exactly the kind of nuanced work that AI is supposed to struggle with.
Will Players Notice the Difference?
That's the real question. AI translation has gotten dramatically better over the past two years. Modern LLMs can handle context, idiomatic expressions, and tone in ways earlier machine translation couldn't.
But there's a difference between "pretty good" and "captures the subtlety of the original writing." Game localization professionals spend years learning how to adapt dialogue, jokes, and cultural references so they land properly in different languages.
An LLM doesn't know that a particular joke won't make sense in Japanese culture. It doesn't understand that certain phrases have different emotional weight in different languages. It just produces statistically probable text.
Online discussion among tech workers suggests this is the beginning of a wave. If game studios can get "good enough" translations from AI at a fraction of the cost, other companies will follow.
The Economics Are Brutal
From a business perspective, this makes sense. Human translators are expensive. AI translation costs pennies per thousand words. If the quality is 85% as good at 1% of the cost, that's an easy decision for a CFO.
But there's a difference between what's economically rational and what's good for the product. Games with bad translations become memes. Players notice when dialogue feels off or when cultural context gets lost.
Warhorse Studios is betting that modern LLMs are good enough that players won't notice the difference. We'll find out when the game launches. If they're right, expect every other studio to follow. If they're wrong, this becomes a cautionary tale about replacing expertise with automation too quickly.
The technology is impressive. The question is whether it's actually ready to replace people who've spent their careers mastering this craft.
