News JVTech “Disrespectful to our profession”: AI robs actors of their voices!
Share :
No artistic field seems to have been spared from the flood of artificial intelligence in recent months. After designers, painters, artists, editors, and developers, it’s the turn of voice actors to stand up to AI “piracy” of their voices.
Artificial intelligence is changing the world, that’s a fact. The whole question is whether the next world will be better or worse than the previous one! One thing is certain, Many professions today are rebelling against the primacy of AI over artistic fieldswhere until recently the human touch seemed inevitable.
The fields of graphics and writing are heavily affected by the use of artificial intelligence, as is the music sector. But Synchronization should also not be neglected. Many players, best known for their activities in the field of video games, are now sounding the alarm.
Movie and game voices soon replaced by AI?
The website motherboard recently gave voice to American voice actors, who find themselves in an uncomfortable position. As a matter of fact, Some of their latest contracts include handing over their voiceprints for use by artificial intelligence in the future.
What does that mean ? Specifically, when a voice actor signs on to dub a character in a video game, they’re recording dialogue for the game in question. After the work is done, the recordings of his voice can then be used to generate further dialogue via an AI.
Fryda Wolff, an actress who is a major contributor to the game Apex LegendsShe’s one of those pros who worry. “Game developers, animation studios, and maybe even commercial clients could get away with draining more power from me by passing my voice to the AI without ever compensating me or even letting me know. my agent. »
For her part, Sarah Elmaleh, known for her voice work on Fourteen days etc gloriole, believes that there may be ethical divergences associated with this approach. “What happens when we take on a role and find in the dressing room that a line of text doesn’t seem correct and makes us uncomfortable? Usually this line can be rejected to prevent its use. This technology bypasses that completely.”
Already drifting into the world of artificial intelligence
Industry professionals point to ambiguous contracts that are not always clear about what they intend to do with the recorded voices. However, it should be emphasized that deviations already exist.
As a matter of fact, Some text-to-speech software like ElevenLabs has recently been in the news for being misused. On 4chan, some community members have used this AI to synthesize the voice of actress Emma Watson (Harry Potter) so she can read excerpts from Mein Kampf.
“It is disrespectful to our profession to suggest that performance is equivalent to real human performance”believes SungWon Cho, a YouTuber specializing in dubbing. “You can make it sound like a voice and maybe even make it sound like it’s capturing an emotion, but it will always end up sounding hollow and wrong. »
Industry veterans like Steve Bloom have taken a public standcredited in recent titles like The Callisto Protocol or midnight suns. “I know that AI technology is exciting, but if you hear my voice or any of the characters I play on any of these sites, please know that I have not given my permission and never will »he wrote on Twitter. “This is completely unethical”.
Hey friends, I know that AI technology is exciting, but if you see my voice or any of the characters that I have featured on any of these sites, please know that I have not given my permission and never will. This is highly unethical. We all appreciate your support. Thank you very much.
— Steve Blum (@blumspew) February 11, 2023
The question is how the industry will position itself in concrete terms and whether it really leaves the professionals a choice: Signing a contract with an “AI clause” or turning down the work could become a difficult decision for voice actors in the future. All of this will be above all a question of rights, because if we know that certain AIs can completely reconstruct a voice with a snippet of just a few seconds, it will probably be impossible to avoid abuse altogether…