When AI recreates the female voice, it also rewrites who gets heard
Voice cloning technology platforms like ElevenLabs allow anyone to replicate a voice using just a few seconds of audio, for a small fee. These technologies are reshaping cultural and artistic expression.
In 2023, Canadian musician Grimes released a clone of her voice, saying that “it’s cool to be fused with a machine”. Similarly, American composer Holly Herndon launched Holly+ in 2021 as a voice tool that sings back music using a “distinctive processed voice”.
These female-led examples demonstrate working with the creative challenges of voice technologies, and in some ways, they’re nothing new: electronic music pioneer and composer Suzanne Ciani developed a technological approach decades ago to incorporate a male persona, named “Steve”, into her compositions when a male voice was required.
Voice-swapping technologies are also used by some male producers to present as female artists. British researcher and musician Helen Reddington has observed that: “Like the male gaze, the male ear is hidden and its power exercised behind the scenes.”
Reddington wrote that in 2018 in relation to the way that male writer/producers use female singers to reach an audience. But applied to AI, this points to a cultural dynamic where voice manipulation, as an extension of the male gaze and ear, may also reflect deeper desires to control female identities – especially in music, where voice is central to emotional expression and identity.
Earlier this month, singer Jorja Smith accused producer Harrison Walker of using AI technology to clone her vocals for his single, I Run. Walker said: “It shouldn’t be any secret that I used AI-assisted processing to transform solely my voice.” But Smith’s record label responded saying the producers producer and his distributors seemed to revel in the resulting public confusion.
AI voice technologies are a form of information technology. However, once a voice is rendered as information and is simulated, it will have “lost” its physical form. The combination of disembodied voice and the quality of its simulation can make it easier for people to think of computers as being like a human. This means that a person will be heard whether a machine or human speaks. It is this connection to a person that voice cloning disrupts.
British AI artist Oliver McCann, known as imoliver, openly admits to having “no musical talent at all … I can’t sing, I can’t play instruments, and I have no musical background at all”. Yet through AI, he has developed songs that foreground a female persona. Likewise, the eminent producer Timbaland has invented a pink-haired female artist, TaTa, in a new genre he refers to as a-pop or artificial pop. But does a creator’s gender matter in the development of AI artists and wider fanbases?
Noonoouri, a digital avatar signed by Warner Music Central Europe in 2024, though not completely AI-generated, is a composite of digital tools, presented as a human-made young female fashionista turned pop star. The Instagram feed for Noonoouri shows everyday activities – eating pasta, throwing peace signs, even signalling support for Black Lives Matter. The avatar’s appearance is malleable. But these gestures and modified appearance, while seemingly empathetic, may be more performative than transformative.
Noonoouri’s creator, Jeorg Zuber, used his own voice – digitally feminised – and motion capture of his movements to animate the avatar. It is Zuber’s embodiment of femininity that is being portrayed here, ultimately to produce a pliable brand ambassador. As Warner’s senior A&R manager, Marec Lerche, has stated: “We can change her style in a minute … we can make her fly if we want.”
As British author Laura Bates, founder of the Everyday Sexism Project, has pointed out, apps used to develop these female avatars and characters rely on a misogynistic idea of “what a woman is and should be. She’ll never disagree with you, she’ll never answer back”. It is not AI doing the impersonation, but the human company.
Technology may blur boundaries, but it also reveals who holds the power. When male creators use AI to simulate female voices and personas, are they expanding artistic possibilities or perpetuating a new form of gender appropriation, ventriloquism and misogyny? One call to action to counter the growth of the manosphere is the increased presence of girl voices to tackle misogyny. Yet voice simulation technologies may undo this.
In the age of AI, impersonation takes on new meaning. When mediated by technologies largely controlled by cisgender men and tech platform companies, female impersonation risks becoming a tool of dominance rather than expression. The question is no longer just about artistic freedom – it’s about who gets to speak, and who is being spoken through.
This article is part of our State of the Arts series. These articles tackle the challenges of the arts and heritage industry – and celebrate the wins, too.
Hussein Boon does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.