According to GameSpot, The Witcher 3 and Cyberpunk 2077 director Konrad Tomaszkiewicz confirmed his new studio Rebel Wolves used AI technologies for their upcoming game The Blood of the Dawnwalker, a 14th-century open-world fantasy RPG where players control Coen, a human by day and vampire by night. The studio specifically used AI to create temporary character voices for early testing before hiring real actors for final recordings. Tomaszkiewicz emphasized he’s “not totally against AI” but believes it should help rather than replace human developers, warning that games made entirely with AI would be soulless. Meanwhile, Elon Musk’s xAI is actively hiring a Video Games Tutor to help develop AI tools for game generation, highlighting the industry’s divided approach to AI integration.
AI as tool, not replacement
Here’s the thing about Tomaszkiewicz’s approach – it’s actually pretty reasonable when you think about it. Using AI for placeholder voice work during development makes perfect sense. You need something to test dialogue flow, timing, and emotional beats long before you’re ready to book expensive recording studios and voice actors. It’s basically like using temporary assets in any other part of game development.
But there’s a huge difference between using AI as a prototyping tool versus trying to replace human creativity entirely. Tomaszkiewicz nailed it when he compared good AI integration to Google Translate – a tool that helps bridge gaps but doesn’t pretend to understand nuance, cultural context, or emotional depth. Can AI really capture the subtle inflections that make a performance memorable? Probably not yet, and maybe never.
The soul problem
When Tomaszkiewicz says AI-only games would be “soulless,” he’s pointing to something fundamental about game development. Great games aren’t just collections of mechanics and assets – they’re expressions of human experience, creativity, and collaboration. Think about his example of finishing the episodic game Dispatch and realizing AI could never create something like that.
And honestly, he’s right. AI can optimize, generate content, and handle repetitive tasks. But can it understand why a particular line of dialogue needs to be delivered with just the right amount of hesitation? Can it feel when a game moment needs to breathe versus when it needs to rush forward? These are human judgments born from experience and emotional intelligence.
Industry divided
The gaming industry’s relationship with AI is all over the map right now. You’ve got Rockstar co-founder Dan Houser saying AI is overrated, Congressman Ro Khanna calling for regulation after Activision used AI for Call of Duty: Black Ops 7, and then Elon Musk going full steam ahead trying to build AI that can generate entire games.
It reminds me of when motion capture first hit the scene – some studios went all-in while others insisted on traditional animation. The ones that succeeded were those who used mocap as a tool rather than a replacement for animator skill. Same principle applies here. The danger isn’t AI itself, but the temptation to use it as a cost-cutting shortcut that sacrifices quality and human touch.
Where this is headed
Look, AI in gaming isn’t going away. The genie’s out of the bottle. But Tomaszkiewicz’s measured approach shows there’s a middle path – using AI for what it’s good at (prototyping, testing, optimization) while keeping humans in the creative driver’s seat. It’s about augmentation, not replacement.
The real test will come when we see the first major releases that heavily rely on AI-generated content. Will players notice the difference? Will they care? My guess is they will – because at the end of the day, we play games to connect with other human experiences, not just consume efficiently generated content. And that’s something no algorithm can fake.
