A short video from 1932 titled Voice transplant is the latest marvel of sound scienceperfects the groundbreaking process by which His Master’s Voice London technicians were able to separate Enrico Caruso’s voice from an old 1905 record. Eleven years after his death, the Neapolitan tenor actually finds himself in a recording studio accompanied by flesh. and a bloody orchestra on the air Wear a jacket come on Clowns of Leoncavallo. “For many years we have tried to bring Caruso’s voice to life, to emphasize its beauty and original dramatic ardor,” explains the commentator: “Finally we succeeded.”
NINETY YEARS the artificial intelligence then resurrects the voices of the missing translators. If the Beatles promise a final reunion with the spirit of Lennon, already awakened in the mid-nineties in free as a bird AND real loveYoutubers form a parallel reality in which Kurt Cobain sings Vanessa Carlton and Freddie Mercury reimagines My heart Will Go On AND I will leave with you. this is the side nasty AI-based speech-to-speech programs like Vocaloid and SoftVC VITS have been on the radar of major companies since at least last April, when Universal forced streaming platforms to remove Heart on my sleeve, the original song of this Ghostwriter-977 (nomen omen) is entrusted to the cloned voices of Drake and The Weeknd. In the expectation that AI can actually compete in musical composition, it is its reproducing nature that raises ethical-legal questions and causes moral panic.
However, Caruso’s example and his autopsy should make us think about a much older problem than the advent of artificial intelligence. A certain necrophiliac flavor, if you think about it, has characterized the recording industry since its inception, if it is true that the famous little dog in front of the phonograph was there to listen to the voice—isolated from the body—of the deceased. owner. As Michel Chion wrote, “those who witnessed the birth of these technologies were well aware of their funerary quality”; I wonder if in the acoustical and schizophonic twenty-first century there is the same awareness of life in the world of the dead who continue to speak (and sing).
Those same posthumous transmission test duets have become happy residential popular music, at least since Nat King Cole was resurrected to sing with his daughter Natalie (Unforgettable, 1991), giving way to a cycle of resurrections. always more ghostly: Lennon, Mercury, Marley, Sinatra, Notorious BIG and the singing company. In 2004 Anita Cochran will publish Changing song compiled from fragments recorded by Conway Twitty, who has long since passed away. Not to mention the holograms: Mercury – still the same – shakes hands with Brian May on stage in Glasgow, and Aiode software resurrects Ofra Haza and Zohar Argov to sing in honor of the 75th anniversary of Israel’s independence.
Practices and technologies that find a common denominator in that recombinant approach characteristic of sound recording and intrinsically dedicated to inter-worldliness: re-arranging a track recorded by a translator who is no longer alive is not much more problematic than conventional multi-track recording is itself chronologically and spatially asymmetric and discontinuous . The will to reassemble what has been removed from synchronization is no different, from the desire between body and voice to the very idea of the voice as a fingerprint.
DEFINITE, the novelty is that AI voice transformation confronts us with performances that can no longer be associated with the human body, alive or not. And, above all, the extraordinary availability of a vilified technological tool is new, which, however, is no more and no less ethical than its users. Even more difficult is the question ofagencyunderstood as the ability to influence the environment, going far beyond the idea of artistic will – both dead and living – and which, even before the advent of artificial intelligence, was also attributed to materials, spaces and technologies.
Among the many exhumations, it is perhaps appropriate to dig out the concept deadness, already outlined in 2010 by Jason Staniek and Benjamin Piekut as a kind of “conditional arrangement” that helps build bio- and necro-worlds that are destined to infiltrate each other. Basically, deadness it is the remodulation of this arrangement in a permanently reversible and recombinant work like the one written down. Who is no longer satisfied with the original idea of preserving and “embalming” the voices of the future deceased, pursuing more and more complex forms of rearticulation of the body inhabiting him.
The editors recommend:
I’ll leave you a song: first of all the entire catalogThese are concepts and approaches that fit well with posthuman theories, from which the relationship between the body and technology is emerging anew after all essential distinctions between bodily existence and computer simulations, biological and cybernetic organisms, robotic teleology and human goals have been abolished. . . .
But at the same time, reading from a Marxist point of view warns us about the idea of necromarketing and about that very late capitalist remodulation between the bio- and necro-worlds. Meanwhile, the record industry, waiting for a decision, preaches badly and raids well, already thinking about how to monetize the nodes of the network where technology and economics, ethical and legal aspects, agency and copyright meet.
THIS IS NOT ACCIDENT it’s possible that Grimes – Elon Musk’s former partner – was the first to take advantage of the vocal cloning market, flaunting an ethic halfway: “I don’t want my voice used to sing the Nazi anthem, unless it’s a joke.”
The editors recommend:
At the Algorithm FactoryThe new ghost tracks, like those that preceded them, are ultimately a great means of bringing past works to life for a modern audience: not only the voice is resurrected, but also its commercial value. Perhaps this is where we should begin to seriously question ourselves by dismissing the discourse on AI in music as part of a broader theoretical and political discussion about new technological revolutions and the “end of work.” Or, rather (or worse), its digitization, which also in the music sector results in the absorption of consumption within production (see Antonio Casilli’s book on this click slaves, published by Feltrinelli), with platforms capable of converting their users’ gestures and choices into unpaid work: even when prosumers upload Kurt Cobain singing A thousand miles Vanessa Carlton, “voice of the master” – quiet, but very lively – sincerely thank you.
In short, there is always the risk of pointing at a straw without noticing the log in your ear. But someone has already prophesied that “in the next two years there will be a battle that we all have to fight: to protect our human capital from artificial intelligence.”
Is this the voice of Marx or Sting?