When Arc Raiders launched last October, it faced an unusual problem for a game that found immediate commercial success: players were upset that characters sounded artificial. The extraction shooter's use of AI-generated dialogue drew immediate criticism despite the game's strong reception otherwise.
Embark Studios' CEO Patrick Söderlund has now walked back the studio's initial commitment to synthetic voices. The studio "re-recorded" some of the AI-generated voice lines in Arc Raiders with human voices, only after its successful launch in October, according to GamesIndustry.biz.
The candid acknowledgment matters. "There is a quality difference. A real professional actor is better than AI; that's just how it is," Söderlund told the outlet. This wasn't a grudging admission. It was a clear statement that the technology had limitations the studio felt compelled to address.
Why the reversal? Arc Raiders' player count peaked at nearly half a million users on Steam, suggesting the game's success happened despite, not because of, its synthetic dialogue. Many players weren't happy with this creative decision even as they kept playing.
The studio's original reasoning for using AI wasn't cynical cost-cutting. Design director Virgil Watkins said text-to-speech acted as "an unlock for us to be able to do voiced characters when we, at the time, did not have capacity to do so". A live-service extraction shooter requires constant iteration. Recording new voice lines for every update, every location ping, every tactical callout would demand repeated studio sessions that might strain smaller development teams.
Yet there's nuance here. Embark Studios paid its actors for approval to license their voices for text-to-speech AI, according to Söderlund. This wasn't voice replacement dressed up as innovation. Actors were compensated for their participation in what remains a complicated middle ground between synthetic and human performance.
A lot of the in-game voice lines in Arc Raiders are now recorded by human actors and there are fewer AI-generated lines in the game than there were when it launched, according to Kotaku's reporting. But it's worth noting the word "some." Söderlund told GamesIndustry.biz that "some" of the AI-generated lines were replaced by voice actors, which could indicate that the studio isn't looking to completely ditch its text-to-speech AI anytime soon.
The team pays voice actors "for the approval to license their voices for text-to-speech for lines that aren't as essential to the immersion of the experience". Ping system audio, tactical callouts, location names: these are the lines that remain AI-generated in many cases. The logic is defensible. Those audio cues repeat constantly and change with every update. They're functional rather than artistic.
What emerges from this is a pragmatic compromise rather than either a complete capitulation to critic pressure or a vindication of AI's capabilities. Söderlund noted that the team uses text-to-speech to "test 15 different lines without recording them, and then we know what to record". The technology serves as an internal tool before human voices seal the final product. That's a more modest claim than either proponents or critics might make.
Söderlund said that the studio pays its voice actors for their time in the recording booth and will "continue to bring many of them back as we carry on updating the game". That ongoing commitment signals that Embark isn't abandoning voice talent as it iterates on its live service. The studio is learning to build differently, not necessarily cheaper.
Arc Raiders' trajectory suggests that player preferences, not technical limitations or cost considerations, ultimately shape these decisions. The game succeeded on its merits despite, arguably because of, its willingness to acknowledge what doesn't work. Whether this becomes industry practice or remains a one-studio course correction remains to be seen. But Söderlund's candour about AI's limitations is itself a step worth acknowledging.