There Is No One Singing This Song
Can a song be emotional if there is no one singing it?
Is emotion just the appearance of feeling?
I can’t think of any expression that is more emotionally powerful than the voice of a singer. When a person writes an effective song with lyrics that are matched by a rhythm and melody that enhance the feeling of the words, the songwriter digs down deep to reflect something real, an emotion that other people can identify with.
Sure, there is bad music, and there are superficial jingles, but when a skilled singer is given an emotionally evocative song, the effect is unrivaled by anything else in human experience.
Instrumental music can be powerful too, but people especially love to hear a singer, because there’s something special in hearing another person express the feeling of a song in such a personal way, in a kind of emotional athleticism, using their own bodies as instruments of something heartfelt.
People go to see live music because the impact of watching a person embodying a bold expression of emotion in real time to a crowd of people who are striving to share in the feeling of the moment together brings about what the sociologist Emile Durkheim referred to as collective effervescence, a communal ebullience. The audience boils over with the singer.
Not all music is live anymore. It’s been generations now that we have possessed the technology of recorded music. Music fans still enjoy listening to recorded music, though most of them will argue that there’s something missing when the performance isn’t live. Some of them seek out a touch of the experience of live music with recordings of live performances.
Music is now taking another big step away from its roots in human-to-human expression with the deployment of generative artificial intelligence tools that can imitate the sound of human musicians, replicating even the voice of a singer.
Generative AI is now capable of composing and creating performances of entire songs in a matter of seconds, a shorter amount of time than what it takes to listen to the audio files that they create. We cannot even call these songs recordings, because they do not record any actual musicians performing. They are pure imitations of musical performance, going straight into an audio file without any physically real performance.
For the first time, we can hear songs that no one has ever performed. We can listen to these songs over and over again without any person ever performing them.
What does it mean for there to be songs that no one has ever sung, songs that no one will ever sing?
Is this an emotional song
if there is no person singing it?
What is a feeling that no one ever felt?
Is emotion just an output, an analytic category
with no humanity required?
There is no feeling in this voice
because there is nobody here.
There is no one singing this song.
It's far more easy to fake an emotion
than to make a heartfelt creation
human to human.
What will this do? What will this do to us?
There is no feeling in this voice
because there is nobody here.
There is no one singing this song.
There is no one here.
There is no one.
No one.
I wrote the lyrics to this song, but I did not perform it. I cut and pasted the lyrics into a generative AI tool that creates audio recordings that imitate the sound of professional musicians, including a voice that sings the words that have been typed into its user interface. This particular song sounds somewhat artificial in its tone, but that’s because it was created with a free version of a generative AI service. Paid memberships in the service create much more realistic sounding songs, including voices that sound like real human singers.
And yet, is the song real, simply because it sounds real? Is it truly a song?
This existence of this song is not just a challenge to the survival of an ancient, uniquely human form of art. It is a challenge to the survival of meaningful emotion itself.
What happens to a society in which computers can create plausible imitations of emotional songs?
The mimicry of emotion is a manipulation.
There is no one singing the song. There is no personality behind the voice. There is nothing but a computer with a sophisticate program that has scanned stolen recordings of actual musical performances, and has analyzed them mathematically in order to reproduce patterns that create the impression of music where no musical performance exists.
That is not the same thing as what happens when a person sings a song.
It is a calculated, emotionless mimicry of emotion. That is the definition of sociopathy.
Music is an emotionally manipulative medium. We agree to listen to it, and we enjoy listening to it, because we appreciate the sincere emotional work that goes into its composition and its performance. A musical performance is an act of devotion to an emotion, and a strenuous sharing of a feeling that comes from lived experience. It’s a meaning that is shared by those who create it with those who listen.
When there is no human effort that goes into the production of music, when there is not even any human composer or performer devoting themselves to the creation and sharing of it, it means nothing.
An artificially generated song asks us to listen without offering anyone to actually listen to.
A technology that can be used to create emotionally-moving media without requiring the work of any human capable of feeling emotion is a tool for the mass production of emotional manipulation.
This technology is currently being offered at a low price, to be tested by the public at large. Behind its current low price, however, is a higher cost of production, a cost that is hidden by the financial backing of venture capitalists who expect to receive a substantial profit, an amount that is much higher than their initial investment. These generative AI features will not remain at their current low price, any more than an introductory rate of Netflix reflects the eventual ongoing cost of membership.
The wealthy people and organizations that own and control the technology of generative AI music imitation will, before long, be able to make the ability to create media of sociopathic emotional manipulation available selectively to other people and organizations that are already wealthy and powerful.
It may sound cute. It may seem like a great toy. Nonetheless, generative AI imitation of music is dangerous. It will be attractive to those who seek power over others without human accountability. It will dull our senses to the special value of human expression of emotion, creating easy dismissal of artistic work, while putting actual human artists, who might be capable of summoning collective resistance, out of work.
A world in which people listen to songs that no one has ever sung is dreary, but worse than that, it is dangerous.
This technology tempts us to forget that emotion is not held in what we hear, or see. Emotion is held in the minds of the people who create songs that we hear, and art that we see.
This technology tempts us to accept the imitation of emotional expression as if it is the same thing as emotion itself.
This technology tempts us out of collective effervescence into collective sociopathy.
Generative AI promises us emotional fulfillment without any intention to actually provide it. Generative AI, after all, is capable of no intention, and no emotion, at all.
Will we listen to this siren song, standing still, enraptured until the very end, until it eats us alive?