A new game seems to have used artificial intelligence text-to-speech instead of human voiceover actors — and man, are the professionals ticked off.

The use of AI speech in "The Finals," a game show-esque first-person shooter created by Embark Studios that's currently in its third and final open beta testing, was pointed out on Twitter by voiceover (VO) actor Gianni Matragrano, who posted a 99-second clip of gameplay that sounded awfully bad.

"So I guess The Finals is going with AI voices...?" tweeted Matragrano, an accomplished voiceover pro who has worked on games ranging from Genshin Impact to Evil West.

The actor proceeded to post a snippet of a July interview on Embark Studios' "Meet the Makers podcast" in which "Finals" audio developers Carl Strandberg and Andreas Almström admitted that they'd used text-to-speech AI for a lot of the game's voices.

"We use AI with a few exceptions," Almström said, "so all the [player] voices like the barks and voiceover commentators are AI text-to-speech."

Notably, most of the game's non-word sounds, such as the grunts and action exclamations uttered by characters portrayed in exertion, were recorded in-house by real people, PC Gamer reports.

"The reason that we go this route is that AI text-to-speech is finally extremely powerful," the audio dev added. "It gets us far enough in terms of quality and allows us to be extremely reactive to new ideas and keep things really, really fresh."

That "really, really fresh" assessment may be a bit of an overstatement. In the clip Matragrano posted, the in-game commentators' voices are stunted and strange, with inflections coming in at bizarre points in their sentences that make them sound, well, robotic.

Almström added that his team preferred using AI to human voices because "if a game designer comes up with a new idea for a game mode, we can have voice-over representing that in just a matter of hours instead of months," but as Matragrano argued, it rarely takes that long for VO actors to get packages done in the industry.

"We are constantly banging out rush order sessions for like, within a day or two," the voice actor said. "We actually make it very easy."

The Genshin Impact actor wasn't the only one perturbed by the AI speech in "The Finals" — or its audio developers' comments — either.

"Why the fuck do AI voice people act like hiring voice actors is some kind of arcane ritual," tweeted Zane Schact, another voice actor. "I've knocked out entire games worth of audio in a two hour session. It ain't deep."

Kit Harrison — another voicover actor from Genshin Impact, not the "Game of Thrones" star — trashed the hypocrisy of using text-to-speech AI for dialogue but not for the exertion sounds that make the game more realistic.

"What really sticks with me is that they needed to bring in real actors to get the grunting, effort, and breathing sounds because the AI can't do it," Harrison tweeted. "It can't replicate the noise that I make when I stand up from my chair, but it wants to take MY job?! Don't make me laugh."

The quality (or lack thereof) of the AI speech in "The Finals" was also pointed out by "Baldur's Gate 3" voiceover actor Elsie Lovelock, who mused on Twitter that "the kicker is, it STILL sounds like crap regardless of how 'realistic' [the developers] think it sounds."

Although this is far from the first game to use AI instead of human actors, this instance seems particularly poignant to those who both work in the industry and play games themselves — not least because the game is still in beta, which should theoretically mean that developers are looking for feedback before a wider release.

In a statement provided to PC Gamer, Embark Studios said that although it used AI voiceovers in the beta versions of "The Finals" and plans to do so in its wider release, "making games without actors isn’t an end goal for Embark and TTS technology has introduced new ways for us to work together."

We've heard the "we don't want to replace humans with AI" line from AI boosters before — and the about-faced statement rarely bodes well for the companies who make them or the people who work for them.

More on crappy AI content: Bing's AI Refuses to Generate Photorealistic Images of Women, Saying They're "Unsafe"


Share This Article