Could AI Craft a GTA 6? Take-Two Boss Thinks Not, and Here's Why I'm Both Skeptical and Intrigued

AI Gaming News Author · Eurogamer ·

Could AI Craft a GTA 6? Take-Two Boss Thinks Not, and Here's Why I'm Both Skeptical and Intrigued

Take-Two boss Strauss Zelnick reckons AI can't create games like *GTA 6*, arguing it lacks creativity and is inherently derivative, posing questions about copyright and true artistic genius. I'm keen on the efficiency AI brings but gently skeptical about its current capacity to replace human vision in crafting truly original, impactful games.

Alright, gather 'round, mates. I’ve been tinkering with some news from the digital ether, and it’s got me pondering the very fabric of game creation. We’re talking about AI, that ever-present whisper in the tech winds, and whether it could ever conjure up something as gloriously chaotic as *Grand Theft Auto 6*.

Our mate Strauss Zelnick, the big boss over at Take-Two Interactive (you know, the folks who publish *GTA*), has weighed in on the matter again. And fair warning: he’s not exactly throwing a parade for our silicon-brained companions. In fact, he reckons AI simply *couldn’t* make a *GTA 6*. His reasoning? "There is no creativity that can exist by definition in any AI model."

Now, that’s a statement, isn’t it? Almost sounds like he’s trying to banish AI to the digital Shadow Realm with a single incantation. Zelnick, as reported by Eurogamer's Matt Wales (and seriously, cheers to Matt for digging into this), has always taken a fairly measured, if not slightly dismissive, stance on AI. He’s called the term "oxymoron" before, which, I gotta admit, has a certain ring to it. He’s "pretty enthusiastic" about AI as a *tool* for efficiency – helping the company do things better, faster. But when it comes to the big 'C' – Creation – he firmly believes that's the domain of "genius," and "genius is the domain of human beings."

His latest comments, dropped during a CNBC Technology Executive Council Summit, really lean into this idea of AI being inherently "derivative." One immediate, and frankly, very real, concern for Zelnick is copyright. Think about it: if an AI slurps up vast datasets of existing content and then spits out something new, whose intellectual property is it? "If you create intellectual property with AI, it's not protectable," he states. And for a company like Take-Two, with multi-billion dollar franchises, protecting their IP is more critical than finding a secret level in a retro arcade game. It's a pragmatic constraint that, let's be honest, is going to get a lot trickier before it gets clearer.

But beyond the legal tangles, Zelnick dives into the philosophical. He argues that AI is inherently "backward-looking." It learns from what already *is*, from old information, and therefore, it can only ever produce something "derivative." It's like trying to predict the future of gaming by only looking at Pong. ("You wouldn't end up with anything very good.") He challenges us: could an AI push a button tomorrow and create an equivalent *Grand Theft Auto* marketing plan? His answer is a resounding "no" on two fronts: it can't *yet*, and even if it could, it would be "pretty derivative." It's good at predictable, data-driven tasks, but anything beyond that? "It's going to be really, really bad at."

And here’s where my digital wizard brain starts sparking. Is creativity *really* impossible for an AI "by definition"? Or is that just a convenient definition to draw a line in the digital sand? I mean, human creativity itself often builds on existing ideas. We remix, we recontextualize, we iterate. Every new gaming genre is arguably derivative of earlier ones – a mutation, a progression. Is the difference that AI does it *mechanistically*, without subjective intent? Perhaps. But then, is intent the sole measure of creativity?

Think about what Rockstar Games achieves with *Grand Theft Auto*. Zelnick calls it creating something that "approaches perfection." That's not just stitching together city blocks or writing dialogue from a database. It’s the intricate world-building, the biting satire, the emergent narratives, the way seemingly disparate elements coalesce into a vibrant, living (if utterly mad) sandbox. It's the 'why' behind the 'what'. An AI could probably generate a million lines of dialogue, but would it land a punchline with the same subversive wit as a *GTA* character? Could it design missions that resonate with the social commentary Rockstar weaves in? That’s where the "genius" argument truly bites.

Now, don't get me wrong, I'm not saying Zelnick is entirely off his rocker. There’s a profound difference between a tool that can generate assets or streamline development processes – which I'm genuinely enthusiastic about, especially for indie devs – and a consciousness that can conjure a cohesive, emotionally resonant, and genuinely *original* artistic vision. AI is brilliant at pattern recognition, at optimisation, at spitting out variations on a theme. And frankly, those capabilities are already revolutionising parts of game development, from procedural generation of environments to intelligent NPCs. It can absolutely help developers "do a better job and to do a more efficient job," as Zelnick himself acknowledges.

But to craft a sprawling, narrative-driven epic like *GTA 6*? A game that often feels like a cultural lightning rod, a reflection and distortion of our own reality? That takes a human touch. It takes lived experience, a sense of irony, a willingness to push boundaries, and sometimes, the sheer bloody-mindedness of a team of people trying to make something truly extraordinary. That's the "genius" Zelnick is talking about, and it's something I reckon our current AI models are still light-years away from replicating.

So, while AI will continue to evolve at a bewildering pace (and honestly, who knows what it'll be capable of next year, let alone next decade?), for now, the idea of an AI-generated *GTA 6* feels a bit like trying to cast a grand illusion with a broken wand. It might look impressive from a distance, but the magic just won't be there. The true alchemy of game development, it seems, still requires a human hand – and probably a fair few late nights fueled by questionable snacks.

**Source:** Adapted from "GTA 6 couldn't be made by AI suggests Take-Two boss: 'There is no creativity that can exist by definition in any AI model'" by Matt Wales, originally published on Eurogamer, October 30, 2025.

Tags: AI in Gaming, Game Development, GTA 6, Take-Two Interactive, Creativity

Original article: Eurogamer