Machines will never take away the voice of the people
By Carlos De la Luz and Le’Raun Peron
In early 2022, a video broadcast on Ukrainian television station, Ukraine 24, showed a cold, emotionless President Volodymyr Zelenskyy ordering his troops to “lay down arms” and surrender to Russia.
The video circulated on social media only to be quickly debunked as a deepfake uploaded by hackers, not only by Ukraine 24 but by people online familiar with Zelenskyy’s cadence who weren’t fooled by the generally unsophisticated video.
Although it ultimately failed to escape the vast uncanny valley, the mere existence of this clip creates certain ominous implications.
“The real threat to me is how it undermines our belief in what is real,” said communication/performing arts professor Dr. Liesel Reinhart. “So the fact that it’s possible to create these dupes gives people the ability to say ‘that’s fake’ when something is real, right?”
“Fake news” could prove to be a far more effective slogan than expected with videos like the one of Zelenskyy mostly serving to bolster the idea that news can be easily manipulated and untrustworthy.
The rapid advancement of AI over the last few years, along with the emergence of notable services like ChatGPT and DALL-E, is bringing this technology closer and closer to replacing human work. One example is the replication of human voices using existing audio recordings. This relatively quick string of advancements is strengthening the potential to easily exploit everyday people as well as artists and other public figures.
“When you record someone’s voice and then have the ability to utilize their voice without them actually speaking the words, it just seems fraught with the possibility of abuse,” Reinhart said.
The ease with which deepfakes can be created in everyday homes has given many the green light to create whatever comes to mind.
Although public figures are easy targets for obvious reasons, they aren’t the only ones affected. Silicon Valley start-up Sanas is working on technology to alter people’s accents. This technology is specifically being utilized for employees at call centers to give the illusion of a Western-sounding voice.
“If that customer is upset about their bill being high or their cable not working or their phone not working or whatever, they’re generally going to be frustrated as soon as they hear an accent,” said Marty Massih Sarim, president of Sanas in an interview with The Guardian.
While the idea of this service may have come from a place of altruism, it inherently sustains ingrained xenophobia in people afraid of foreign accents, as well as essentially erasing part of the original speaker’s identity. The human voice doesn’t define a person, but it shouldn’t be altered to cater to prejudice.
“The voice is a representation of the person, their beliefs, their identity,” Reinhart said. “And so it’s not just a wave file; a voice is more than a wave.”
The art world, in particular, has had a very tumultuous relationship with AI. Whether it’s filmmaking or music, the human voice shining through is the key component in any piece.
This last year saw two of the longest labor strikes in Hollywood history, with the WGA and SAG-AFTRA strikes. The writers and actors both dealt with the penny-pinching studios’ refusal to come to a fast agreement where a huge point of contention was the exploitation of artists through AI.
The people running these studios are completely aware of the potential AI has in saving them a considerable amount of money. If an actor delivers a certain line of dialogue and the decision is made that they should say something else instead, they would traditionally be brought back to re-record the line. With AI, the studio could avoid having to pay the actor for more work and instead just use pre-existing audio they already have.
“That is one of the things that is a great thing about art is it communicates those things which are not communicable otherwise, right?” said PCC Center for the Arts manager Zak Graff. “Whether that’s the heart, the mind, or the soul, that’s open for interpretation. But that’s how people receive and contemplate art.”
Save for voice actors, an actor’s most valuable asset is their face, to be as superficial as possible. With singers, their success is dependent almost entirely on their vocal skills as well as their distinct personalities.
“I think each one of us are so special individually as musicians that it’s really weird when they start just producing stuff off of somebody’s voice,” said PCC jazz voice instructor Mary Durst.
Like any art form, music is an expression of emotions and ideas from a real person. Artists can use technology as tools to enhance their craft, but those tools are simply that and don’t have the emotions necessary to create.
“I think that technology and electronics have already influenced music in a substantial way, but so far, it hasn’t been replicable without a human component,” said music performance major Hannah Stubbs.
The more we look, the more we see it, the more we watch, the more we see it, and it’s now at the point where the more we listen, the more we hear it. But will we even know when it comes across our ears?
“So I think that there is going to be a noticeable difference between humans and AI, but it’s already very advanced,” Stubbs said. “So continuing to advance without regulation, it could pose a threat to musicians.”
There are a lot of different aspects that can be taken into consideration regarding AI in today’s music. One that seems to be the most fresh and potent is the replication of current artists’ music, releasing it as if it’s authentic and taking it to places where it doesn’t belong. The most recent example of this is Drake and The Weeknd with the fake song “Heart on My Sleeve.”
Now, most people in the music loop would know that the two Toronto legends have no plans of collaborating on any music in the near future, but seeing how far this fake song got is scary enough. And to have two of the biggest artists named and it still gets that far, what message does that say to other songwriters and producers? Nobody is safe?
Berklee Online breaks down each level of the ethical issues this can and will cause. Writer Talia Smith-Muller does highlight that these actions will be followed with consequences, but that’s only going to apply to the ones that are caught and don’t slip through the cracks.
Another aspect that can’t be looked over is the “music” that will be mimicked from artists who are no longer alive. Young and old singers and rappers alike can both run into this dilemma, and with the main source not here to speak for themselves, who can tell the audience what is and isn’t real?
GQ Magazine and Billboard have both covered these situations with different examples that have transpired. GQ writer Daisy Jones goes in on this practice calling it “exploitative and tacky as hell.”
Billboard notes how a Chinese company named “Tencent Music Entertainment” takes part in this and has a song that currently earned over 100 million streams.
Now depending on the fanbases of the artists that are being mimicked, they may or may not have a problem with this as a whole. On social media which is used by the younger side, these little mix and mashups are enjoyable and satisfy the itch of curiosity in their brains.
However, for the older generation, there is a “growing anxiety” due to the endless possibilities of what this can mean for the music industry in the future as Laura Molloy explains in “Should we use AI to bring musicians back from the dead?”
Everyone involved in any form of media knows and can see how artificial intelligence is slowly creeping up and taking over socially. As of right now, it’s all fun and games to pick and point out where it’s obviously not a real person talking or singing. But when fake pieces of work start to fall through the cracks and get by without being noticed, who will be laughing then?
