According to GameSpot, former Subnautica director Charlie Cleveland, Unknown Worlds co-founder Max McGuire, and ex-CEO Ted Gill have filed new court documents alleging Krafton executives consulted ChatGPT about avoiding a $250 million performance bonus. The documents claim Krafton CEO Kim Chang-han and global operations head Richard Yoon discussed using AI to brainstorm ways to cancel the earnout payment tied to Subnautica 2’s financial performance. Messages between executives allegedly show Yoon suggesting “it might be easier to take over [Unknown Worlds]” than renegotiate terms. Krafton now claims the founders downloaded confidential files, though this justification emerged after their termination. Subnautica 2 is still expected to hit early access in 2026 on PC and Xbox Series X|S.
The AI legal strategy backfires
Here’s the thing about using AI for legal strategy—it creates a paper trail. Or in this case, a digital one that apparently doesn’t exist anymore. The former developers allege Krafton refused to produce those ChatGPT conversations and claims they no longer exist. That’s convenient, isn’t it? But the fact that they’re even discussing AI consultation in internal messages suggests how seriously they were considering every possible angle to avoid that massive payout.
Krafton’s shifting legal ground
Krafton’s legal position has been all over the place. They initially accused the founders of planning to rush Subnautica 2 into early access to trigger the bonus—then withdrew that claim when pressed for evidence. Now they’re saying the founders took confidential files, which somehow justifies a termination that happened before this alleged misconduct was discovered. It feels like they’re throwing everything at the wall to see what sticks. And honestly, that rarely looks good in court.
The “AI-first” culture clash
Krafton recently declared itself an “AI-first company,” which immediately prompted Unknown Worlds to publicly state that Subnautica 2 won’t use generative AI. That’s not just a technical disagreement—it’s a fundamental philosophical divide about game development. When your publisher is consulting ChatGPT about bonus structures while you’re trying to create an immersive underwater survival experience, there’s clearly a cultural mismatch happening.
Broader implications for game dev
This case could set some worrying precedents. If publishers can use AI to find loopholes in performance bonuses, then fire developers before those bonuses trigger, what does that mean for the entire industry? Performance-based compensation has always been tricky in game development, but adding AI into the mix creates entirely new ethical questions. Basically, we’re watching a high-stakes test case for how AI will be weaponized in corporate disputes.
