Stock image of Australian software developer Replica Studios’ voice synthesiser platform for games developers

When the epic open-world PlayStation 4 game Red Dead Redemption 2 was developed in 2013, it took 2,200 days to record the 1,200 voices in the game with 700 voice actors, who recited the 500,000 lines of dialogue.

It was a massive feat that is nearly impossible for any other studio to replicate – let alone a games studio smaller than Rockstar Games.

But with advances in artificial intelligence it is becoming easier and easier to recreate human voices to create automated real-time responses, near limitless dialogue options and speech tailored to a user’s unique input. But the technology raises questions about the ethics of synthesising voices.

The Australian software developer Replica Studios rolled out a voice synthesiser platform for games developers in 2019 – a tool used by Australian games developer PlaySide Studios in their game Age of Darkness: Final Stand.

“We would hope that there’d be hundreds, if not thousands of other studios that could dream of building games like [Red Dead Redemption 2] because everyone wants to do that,” Shreyas Nivas, the chief executive of Replica Studios, says.

Recording every line of dialogue individually is “so inefficient from a cost perspective, but also from a time perspective, and you need to have these huge teams”, Nivas says.

Replica has licensed the voices of 120 actors for use in video games, which are capable of up to 1,000 different vocal tones, according to the company.

Nivas says he sees AI voice synthesising as the future, but, as with many AI advances, the practice is fraught with ethical dilemmas.

There are now free online voice synthesiser tools that can be used to mimic celebrities’ or film and TV character voices – often without the permission of those artists. And Bloomberg reported this month that some voice actors were “shocked” to discover their voice being used in content they had not participated in. (They later learned their contract terms had been broad enough to cover such uses.)

In Japan earlier this month, the Japan Performing Arts Workers’ Association held a press conference raising concerns about the impact AI will have on voice acting and music. The group – which reportedly has about 52,000 workers as members – called for legislation to protect their jobs.

Nivas says the licensing model Replica has pursued allows actors to keep earning from the use of their voice even when they’re not in a studio recording.

“There’s a lot of things to get right, but we’re trying to build this in a way in which we take the voice actors along on a journey with us. So we’re trying to be the most transparent about the process.”

Stock image of Replica Studios’ voice synthesiser platform. The software developer believes AI will give smaller games studios easier access to voices – and early-career actors more work. Photograph: Replica Studios

Nivas says the bigger video game companies will probably still employ high-profile talent, like Troy Baker who voices Joel Miller in The Last of Us, but AI will give smaller games studios easier access to voices – and give early-career actors more work.

skip past newsletter promotion

The company has also implemented content moderation policies for the use of voices in scams or illicit content. Nivas says it is still a learning process as AI accelerates, but he believes the acceleration should continue.

“There’s a lot of this where we don’t know the outcome, but the best we can do is be wary of how we go forward. But keep going forward.”

The Australian union that represents actors, the Media, Entertainment and Arts Alliance, has yet to formulate a policy focused specifically on AI’s impact on acting but a spokesperson told Guardian Australia the promise of AI should not come at the cost of quality, meaningful work for performers.

“Actors must retain ownership of their voices and this should not be given away through a licence in perpetuity,” the spokesperson said.

“In this emerging field of technology, we urge producers to engage with performers’ representatives.”

The Canadian singer Grimes last month invited people to create new songs using her voice synthesised with AI, and said she would split the royalties 50-50.

Earlier this week she tweeted she was stressed that people are “starting to make competitively (or maybe better??) quality grimes sounding songs than I do”.

“But it’s also the most wonderfully poetic way to die and respawn in another career.”

I am actually kinda stressed that ppl r starting to make competitively (or maybe better??) quality grimes sounding songs than I do but it’s also the most wonderfully poetic way to die and respawn in another career

— 𝔊𝔯𝔦𝔪𝔢𝔰 (@Grimezsz) May 8, 2023


Share This Article