The Finals is a multiplayer shooter with a bit of buzz about it, fresh off an open beta weekend on PS5. Some punters, however, were quick to point out that the in-game voices felt a little slapdash; as it turns out, they were generated using AI, forgoing the use of traditional human actors in favour of something more algorithm-based.
Embark Studios designer Andreas Almström confirmed the use of AI text-to-speech technology in an interview back in July, which was just recently picked up by voice actor Gianni Matragano. Appearing on Embark Studio's own Meet the Makers podcast, Almström was fed a fake question, "Who did the voiceovers? They sound really authentic", to which he proudly replied: "So here’s the kicker: what did the voiceovers?"
Almström elaborated, stating: "The thing is, we used AI with a few exceptions. All the contestant voices, like the barks, and both of our commentators are AI text-to-speech. Things we call vocalisations, we can’t really get the AI to perform those kinds of tasks, yet." These "vocalisations" are the breathy grunts a character makes when performing actions like running, vaulting, or jumping. This work was done in-house by Embark Studios developers.
The use of AI anything is a pretty hot-button issue in creative industries like video games, where the livelihoods of human beings are directly impacted by the inevitable race to the bottom. Almström explains that: "The reason that we go this route is that AI text-to-speech is finally extremely powerful. It gets us far enough in terms of quality and allows us to be extremely reactive to new ideas and keeping things really, really fresh.”
One of the supposed upsides of this kind of AI technology is that it's quick and cheap. Matragano pushes back on that thread, explaining that professional voice actors, post-COVID, are often capable of recording high-quality audio remotely and turning around a job inside a day or two: “You can literally get pro-grade VO for less than a grand total, bang out a couple recording sessions and bam you have all the audio you need. We actually make it very easy. And then it'll just sound good and not be something even players who don't really care about AI ethics keep complaining about.”