Even after its success, Embark design lead is still lukewarm on AI usage for Arc Raiders: ‘I don’t think it’s fallen any way or the other’

Arc Raiders was one of 2025’s hits—and while it’s always hard to predict which live service will be the next actually successful one, that by itself isn’t the most notable thing about it. It’s an extraction shooter, and if you get one of those off the ground, it’ll continue to be popular. Its largest impact on the industry has been the conversation around AI.

The game’s developed by Embark Studios, creators of The Finals, and has routinely dipped its toes into hot water over the usage of generative AI to produce voice lines. Statements have sometimes even contradicted each other.

For example, in the story I just linked, Embark Studios CEO Patrick Söderlund was quick to emphasise that the studio wasn’t using it to “replace people”, even though replacing work you’d otherwise need a voice actor to come in and record is, uh, doing that. He then goes on to talk effusively about how Embark couldn’t have made its two games without it, which strikes me as strange—either you replaced work that would’ve cost money you didn’t have, or you didn’t do that.

I digress: You might think that after gathering some success, Embark would keep banging that drum. But in a recent interview with PCGN, design director Virgil Watkins seems pretty lukewarm. Neutral. Straight down the middle.

“Honestly, I don’t think it’s fallen any way or the other,” Watkins says, after being asked whether Embark was all-in on AI now that Arc Raiders is the new hotness. He explains that it’s a case of: “‘Does it ultimately let us do something we couldn’t before, or is it an added [bonus] to the game?’ With the [text-to-speech] stuff, I think it was an unlock for us to be able to do voiced characters when we, at the time, did not have the capacity to do so.

“Do we have different affordances, now that the game is what it is? Probably. And then we can ask ourselves, ‘Did the quality hit the mark?’ And maybe not.”

That’s a pretty big admission that the game’s text-to-speech AI generation is a stopgap, especially since its design director is iffy on the result being high-quality. Which it isn’t; I’ll confess, I bounced off Arc Raiders—partially because I don’t like extraction shooters, but also because I’m a curmudgeon when it comes to AI, and hearing the listless, bland text-to-speech rattle off voice-lines made me feel like I’d been tricked into watching YouTube shorts slop, not playing a videogame that costs money.

Essentially, Watkins says that it’s not “been like, ‘Oh, well, let’s open the floodgates for all types of AI or even AI-adjacent tools.'” However, the backlash hasn’t given Embark much pause on how it’s been using generative AI, either.

“It is still very much in the vein of building what we can, the best we can. And a lot of it is just that avenue of exploring emerging tech and building our own tools for things, because that’s what enabled us to build a lot of this with the team the size we have.

“So I think it’ll be more of that in the future, and just trying to see what we can do for ourselves to, like, keep building content at the scale we have. But obviously we’re not deaf to the concerns that are out there for it.”

As our own Mollie Taylor observed last month, 2025 really has been the year of studios’ limit-testing AI usage, and while Arc Raiders’ success is proof it’s naive to expect people to vote with their wallets en masse, it’s interesting that even a studio that found success with the tech is still kinda ambivalent on it.

And sure, small studios, big dreams, couldn’t afford voice actors, yadda yadda yadda. I don’t know how much I’m convinced by all of that, even if I can see the argument—but I’m more or less with Neil Newbon on this one. If you’ve got the cash now, surely paying actual voice actors is pocket change.

About