Arc Raiders' audio director argues in favor of its AI text-to-speech, saying that "actors would just go crazy if they had to read all of that."
Arc Raiders came under fire last year by using AI text-to-speech for NPC characters found throughout the game, with Embark Studios co-founder Stefan Strandberg saying "there's no end goal in replacing any actors," and that the team used it to "maximize what we can, being a small team trying to get make rich worlds right, but still staying small."
Meanwhile, the CEO of Nexon said we should "assume that every game company is now using AI." Naturally, this caused backlash from players and developers, as these minor NPC roles were ones that could have been filled by professional actors, like NPC roles in countless other games.
In an episode of Embark's "The Evolution of ARC Raiders" video series (via Eurogamer) focusing on sound and music, audio director Bence Pajor recalls previous experiences in development when teams were unable to change gameplay decisions, calling it a "true pain." Pajor explains that while the training data for the text-to-speech was made with actor's performances, he argues that it wouldn't be feasible to have them record all of the lines, saying "the actors would just go crazy if they had to read all of that."
Diana Gardner who plays Celeste in the game, called the tech "freeing" as "I can go and take a job somewhere else," and if Embark needs her input, it'd be short enough that she doesn't have to cancel other gigs.
Despite arguments that using AI for performances does just that, Embark CCO Stefan Strandberg says "we are not in the industry of replacing actors," saying what the developers are doing is "the opposite." He explains, "We should celebrate the performance and give them more space. That's something I'm hoping to give to players over the coming years."
