Game bosses say AI is making studios more creative. The people making the games disagree.

Gaming 4 min 80 sources

A week of dueling accounts — executives at EA and Epic praise AI's "rise in creativity" while developers, a union, and even a former AI chief say the opposite — set against fresh layoffs, a record GTA 6 price, and indie breakouts the giants didn't see coming.

Key takeaways

Game executives at EA and Epic say AI is making studios more creative; developers, a union, and even a former AI chief say the opposite — and which story you believe depends on where you sit.
The industry's layoff summer rolls on, with EA cutting staff for the third time this year, acclaimed studios closing, and a French union calling a sector-wide strike.
GTA 6 set a new $80 price and shipped disc-first as digital-only — while the year's real breakout was a seven-million-selling indie the giants didn't make.

This week the games industry told two stories about itself at once, and they don’t match. Executives say artificial intelligence is making their studios more creative. The people who actually make the games say it isn’t. Both can’t be right — and which version you believe depends mostly on where you’re standing.

The view from the top floor

On Wednesday, EA’s President of Enterprise Development, Laura Miele, said generative AI — software that can produce text, images, or code on command — is driving “faster prototyping” and “a real rise in creativity” inside the company’s studios [7][38][41]. Epic said AI will play a bigger role in the next version of Unreal Engine, the toolkit much of the industry builds on [22]. The framing from the executive floor is consistent: AI is a productivity gain, and productivity is good.

It lands oddly, though. EA said this in the same stretch it ran its third round of layoffs this year — cutting recruitment, customer support, trust-and-safety, and IT staff, including people at its Hyderabad office in India who’d been there more than ten years [6][8]. And it comes ahead of EA’s roughly $55 billion sale to a group led by Saudi Arabia’s wealth fund [5][8]. When a company praises a tool that makes work “faster” while shrinking its workforce, developers hear the quiet part: faster can mean fewer.

The view from the desk

The people doing the work tell it differently. John Buckley of Pocketpair, the studio behind the hit Palworld, was blunt: “Gamers don’t want it” — and so the studio doesn’t touch generative AI [27]. A former Dragon Age writer warned that AI could make game development “frustrating as hell,” and asked the question executives skip: “How are we going to train up the next generation of devs if we eliminate every entry-level task?” [64]. The first rung of a career is where people learn the craft. Saw it off, and the ladder still stands — but no one new can climb it.

The sharpest dissent came from inside the executive world itself. Take-Two’s former head of AI — the company that makes GTA — said the generative-AI hype is “poisoning the well,” and could sour the industry on even the older, narrower kinds of AI that already work quietly inside games [28]. When the person hired to champion the technology says the marketing is the problem, the gap between the two stories isn’t about the tool. It’s about who’s describing it.

A summer that keeps cutting

The dueling AI claims sit on top of a grim labour backdrop. France’s independent video-game union, the STJV, called for “the entire video game sector” to strike on June 25, citing relentless layoffs and bosses who “continue to lounge in their mansions with indoor pools, even if their studios are shut down” [1]. The closures it’s protesting are real and specific: Douze Dixièmes, the French studio behind the acclaimed early-2026 game MIO: Memories in Orbit, is shutting down months after release [16]; the team behind the shooter Luna Abyss was laid off less than a month after the game shipped [11]. A studio can make something good and still not survive the year — and the founder of Eidos-Montréal put the reason plainly, saying modern executives are driven “more by spreadsheets than passion for games” [37].

GTA 6 names its price — $80, no disc

Into that contraction, the industry’s biggest title drew a new line. Take-Two priced Grand Theft Auto VI at $79.99, breaking the $69.99 ceiling that top games held for years, and added a $100 “Ultimate Edition” on top [12][44]. Analysts framed the move as permission for the rest of the industry: if the biggest game can charge $80, the next one can too [23]. The launch is digital-only at first — no physical disc — and at least two retailers said they won’t stock it until there’s a disc to sell [26]. The friction is the point: a game you buy without a disc is a game you can’t resell or lend, and a price floor one studio sets becomes the floor everyone else stands on.

What the giants didn’t make

The quiet good news ran underneath all of it. While the majors cut staff and raised prices, players spent record amounts elsewhere — global games revenue passed $200 billion, with PC leading the growth [3]. And the year’s surprise hit came from outside the blockbuster machine: Meccha Chameleon, an artful hide-and-seek game, sold around seven million copies in two weeks [4]. Polygon noted the fiercest Game of the Year race is among independent games, not the giants [21]. The studios with the most spreadsheets aren’t the ones making the games people can’t stop talking about — which is its own answer to the question of where creativity actually lives.

02 · Lesson · why it matters

Whose report of the work do you believe?

When two honest people describe the same thing in opposite ways, the disagreement usually isn't about the thing — it's about which seat they're sitting in.

Two true sentences that contradict each other

An executive says the new tool made the studio more creative. A developer at the same studio says it made the work worse. Both are reporting what they see. Neither is lying. And they cannot both be describing the same reality — unless they aren’t.

This week that split ran right through the games industry. EA’s leadership praised artificial intelligence for a “real rise in creativity.” The people making the games — a Palworld developer, a former Dragon Age writer, even the company that makes GTA’s own former AI chief — said something closer to the reverse. The instinct is to ask who’s right. The more useful question is: what does each of them actually have in front of them?

The view from the spreadsheet

The executive isn’t looking at the game. They’re looking at a dashboard. Prototypes shipped per month. Hours from idea to playable demo. Cost per task. From that seat, a tool that turns three days of work into one looks like nothing but gain — the number went up, and the number is the whole view.

This isn’t a story about a bad person. It’s a story about a vantage point. A spreadsheet is built to make some things visible and everything else disappear. It can count prototypes; it cannot see whether the prototype is any good, whether the person who made it learned anything, or whether the speed came at the cost of the next ten people who’ll never get to make the slow early version themselves. The dashboard isn’t lying. It’s just answering the only question it was built to answer — and then someone reads the answer as if it were the whole truth.

The view from the desk

The developer is sitting somewhere the dashboard can’t reach. They feel the texture of the work — the part where you wrestle a problem until it gives, and the wrestling is where the idea actually arrives. The former Dragon Age writer named the piece that no metric catches: if you automate every entry-level task, “how are we going to train up the next generation of devs?” The first rung of the craft is invisible from the top floor, because from up there it just looks like slow, expensive work being made fast and cheap.

So the developer’s “this is worse” and the executive’s “this is better” are measuring different things and calling them by the same word. Creativity on the dashboard means output. Creativity at the desk means the thing you can only reach by going through the hard middle. The word survives the trip between floors; the meaning doesn’t.

Why the higher seat usually wins the argument

Here’s the part that matters past gaming. When two accounts conflict, we don’t usually weigh them evenly. The account that comes with a number tends to win — because a number sounds like a fact and a feeling sounds like a complaint. “Prototyping is 40% faster” beats “the work feels hollow now” in almost every meeting, every quarterly report, every headline.

But the number isn’t more true. It’s more legible — easier to put in a slide, harder to argue with, built to travel. The feeling at the desk might be the more accurate report of what’s actually happening; it just can’t be put in a chart, so it loses the argument to a thing that can. That’s why the former AI executive’s warning — that the hype is “poisoning the well” — was so striking. It was a person from the spreadsheet floor saying: the spreadsheet is mismeasuring this. The dissent only counted because it came from a seat the room already trusted.

You are sitting somewhere too

It’s tempting to read this as a lesson about them — bosses fooled by their own dashboards. But the trap isn’t a flaw in executives. It’s a flaw in seats. Every one of us judges the world from a vantage point that makes some things countable and the rest invisible, and then mistakes the countable part for the whole. The manager can’t see the craft. But the developer can’t see the payroll math, or the investor who’ll pull funding if prototypes slow down. Each seat is honest and each seat is partial.

The reader is in a seat too — looking at this very industry through sales figures and review scores and headlines, none of which can show you the person who lost a ten-year job at an office in Hyderabad, or the studio that made something lovely and closed anyway. When you next meet two accounts of the same thing that won’t reconcile, the humble move isn’t to pick the one with the number. It’s to ask what each seat can see, what each one can’t, and to hold your own conclusion a little more loosely — because you, too, are reading off a dashboard, and the dashboard was never the world.

03 · Lab · your turn

Two Seats, One Studio

Read the same studio events from the executive's dashboard and the developer's desk, and feel the identical facts flip between win and loss depending on what each seat can see and can't.

04 · Hope · carry this

The thing worth making was never something a dashboard could count — which is why no spreadsheet can take it from the people who do the work. This week, a small team that simply made something good outsold the giants.