Gina's story:

Gina's story is one of replication and commodification of the human experience. In 2017, Gina, an American voice artist who lives and works in Brussels went for an audition to be the voice of a voice enabled digital assistant. The audition process took days as their were thousands of short scripts to read. Gina didn't get the gig but she was paid the industry award hourly rate for her time.

Two years later, while back in the United States visiting her family, Gina was riding in a taxi when she recognised the voice of the GPS navigator as a slightly disjointed version of her own voice. After some investigation it became clear that the recordings made earlier in Brussels had been sold onto a third party and that numerous examples of unauthorised reproductions of her voice existed on the internet.

Gina wasn't able to claim royalties for any of the unauthorised incidents of her voice on the internet as it is technically no longer her voice. Spectrograms of Gina's early recordings were spliced into hundreds of thousands of smaller spectrograms, each representing different phonemes which were then used to reconstruct the sentences used in the voice overs.

This story provides an unusual example of a person becoming aware they have been commodified in the production cycle of an AI technology. However, replication and appropriation of human characteristics and the experience of being of human are an intrinsic component of how many AI driven technologies insert themselves into the human social world.

My name is Gina, I am a voice artist. In 2018 I was in the United States visiting my family, when I was invited to audition for the voice of a digital assistant. There were lots of scripts to read, varying from personal declarations to observations about the finance market.

I didn’t get the gig but I was well paid for my time. Two years later, I was in a New York taxi, on my way to meet a friend, when the driver switched on the navigator, and I heard my voice say “left at fifth and twenty third street”.
After some investigation, I discovered that the audition tapes I had made earlier, had been converted into thousands
of spectrograms, each representing a different sound in the English language, and that Text to Speech technology had been used to recompose those spectrograms to create new words, ones I hadn’t recorded but in a voice indistinguishable from my own.

Since then, I have found examples of my voice advertising products on YouTube. Legal advice I sought assured me that I am not eligible for royalties as the process of splicing and recompiling of the audio files mean that it is no longer technically, my voice.