AI-Generated Music vs. Human Artists: What Is at Risk?

Why human creativity, artistic identity, transparency, and the connection between musicians and listeners matter more than ever.

In July 2026, Deezer published a number that should make everyone who cares about music stop and think.

At peak levels in June, more than half of all new tracks delivered to the streaming platform were fully AI-generated. Deezer reported receiving approximately 90,000 AI-generated tracks per day. It also found that up to 85 percent of the streams generated by fully AI-created tracks in 2025 were fraudulent.

These figures come directly from Deezer’s report on the rapid growth of AI-generated music.

This does not mean that half of the music people currently listen to is AI-generated. Deezer says fully AI-created tracks still account for only around 1 to 3 percent of total streams on its platform, partly because detected tracks are excluded from algorithmic recommendations and editorial playlists.

Nevertheless, the enormous volume of new uploads reveals how quickly streaming catalogs can be flooded.

This is no longer a theoretical discussion about what artificial intelligence might one day do to music.

It is already happening.

Music can now be generated in seconds, distributed automatically, assigned to invented artist identities, and uploaded on an enormous scale. The barrier to producing something that sounds polished and convincing has almost disappeared.

As an artist who has spent more than two decades composing, producing, collaborating, and releasing music—and as the author of From Silence to Sound—I find this development deeply troubling.

Not because I am against technology.

But because I believe technology should support human creativity, not replace it, imitate it without permission, or make human artists invisible beneath an endless stream of generated content.

At the same time, I do not believe human creativity will disappear.

On the contrary: the more synthetic content surrounds us, the more valuable genuine artistic identity may become.

I Am Not Against AI

It is important to make this clear from the beginning.

I use AI myself, and I appreciate many of its possibilities. It can assist with research, organization, translation, technical questions, and creative exploration. It can offer a different perspective when you feel stuck.

During one production, for example, I used AI to suggest an alternative chord progression. It helped me break out of a loop and look at the piece from another angle.

But the melody, arrangement, sound design, performance, production decisions, and emotional direction remained mine.

That distinction matters.

There is a fundamental difference between using AI as a supporting tool within a human creative process and asking a system to generate an entire track that is then presented and monetized as someone’s artistic work.

AI use exists on a spectrum.

A producer might use an intelligent tool to clean up a recording, analyze a mix, organize files, suggest an idea, or complete a repetitive technical task. In these cases, the human artist still defines the intention and makes the meaningful creative decisions.

At the other end of the spectrum, someone may enter a prompt, generate an entire song, create an artificial artist identity, and upload hundreds of similar tracks without clearly informing listeners.

Those situations should not be treated as the same thing.

This is why a broad coalition of music organizations has proposed separate labels for AI-Assisted and AI-Generated recordings. The aim is to help listeners understand whether AI supported a substantially human creation or generated the primary creative elements itself. You can read more in the IFPI announcement on AI music labeling.

A synthesizer does not compose a piece of music for me. A reverb does not decide which emotion I want to express. My digital audio workstation does not determine what the track means.

These tools expand what I can do, but the decisions still come from me.

AI becomes something fundamentally different when it replaces those decisions.

The Problem Is Not Experimentation

I do not want to exclude anyone from making music.

You do not need decades of experience, expensive equipment, formal education, or professional success to be an artist. A beginner writing an honest first song can be an artist. Someone working entirely on a laptop can be an artist. Someone using samples, generative tools, unconventional instruments, or AI assistance can be an artist.

Experience does not determine whether someone is allowed to create.

What matters is whether a person is genuinely involved in the creative decisions, has something of their own to express, and takes responsibility for what is presented as their work.

The problem is therefore not that technology makes music production more accessible. Accessibility can be a wonderful thing.

The problem begins when accessibility turns into automation without responsibility.

One AI-generated track will not destroy music. Curiosity and experimentation have always been part of creative culture.

The real concern is what happens when generative AI is combined with mass distribution, misleading identities, automated uploads, streaming manipulation, and a business model based on volume rather than artistic value.

A human artist may spend weeks or months developing a single track. An automated system can produce hundreds or thousands of tracks during the same period.

Those tracks can be distributed under numerous names, optimized for particular moods or search terms, and uploaded in the hope that a small percentage of them will generate streams.

Instead of creating one piece of music and believing in it, the strategy becomes generating as much content as possible and waiting for the system to reward something.

Spotify reported removing more than 75 million spammy tracks within twelve months during a period marked by the rapid expansion of generative AI. Spotify did not say that all of these tracks were AI-generated. However, it warned that AI has made tactics such as mass uploads, duplicates, artificial track-length manipulation, and other forms of music spam easier to exploit.

You can read Spotify’s explanation in its announcement about stronger AI protections for artists, songwriters, and producers.

AI is not the only problem.

It is the accelerator.

The deeper problem is a system that can reward scale, repetition, deception, and automated consumption more easily than patience, originality, and artistic development.

When an Artificial Artist Looks Real

The debate became especially visible in 2025 with the appearance of The Velvet Sundown.

The supposed rock band accumulated more than one million streams on Spotify. It had artist photographs, a biography, a recognizable visual identity, and several releases.

But the musicians shown in the images did not exist.

The project eventually acknowledged that its music, imagery, and fictional background had been created using generative AI under human direction. The case led to renewed calls for streaming platforms to identify AI-generated music more clearly.

The Guardian reported on The Velvet Sundown and the debate surrounding its success.

The problem is not necessarily that someone created an experimental AI project.

The problem is that many listeners initially had no clear way to understand what they were supporting.

Was this a group of musicians building a career together?

Were the photographs real?

Were the voices connected to identifiable performers?

Was there an artistic history behind the project?

Or was the entire identity a synthetic construction?

Listeners deserve to know.

Transparency does not prevent anyone from listening to AI-generated music. It simply allows people to make an informed choice.

What Does It Mean to Be a Human Artist?

Being an artist is not about fame, commercial success, formal education, or technical perfection.

It is about intention.

A human artist brings experience, memory, taste, curiosity, vulnerability, and personal history into the creative process.

Why this chord and not another?

Why should the arrangement remain empty here?

Why should the melody rise at this moment?

Why does one sound connect to a memory while another does not?

Which imperfection should remain because correcting it would also remove its emotional character?

These decisions grow from a person’s life.

They also develop over time.

I started making music as a child. Across decades of composing, producing, collaborating, and releasing music, my artistic identity did not appear in a single moment.

It developed through experiments that succeeded and others that failed. It grew through unfinished ideas, difficult decisions, changing technology, collaboration, self-doubt, and the gradual discovery of what felt honest to me.

Every album became part of that journey.

In my own ambient and electronic music, a large part of the creative process is often deciding what not to add.

A sound may appear only once. A harmony may unfold slowly over several minutes. A moment of silence may carry more meaning than another layer. A subtle change in texture may matter more than an obvious climax.

These decisions are connected to instinct, memory, emotion, and years of listening.

They cannot be measured only by how polished the finished track sounds.

This is also why I wrote From Silence to Sound – Unlocking Creativity in Music Production. The book is not only about operating tools. It is about developing an artistic voice, trusting your instincts, overcoming creative obstacles, and using technology without losing your identity.

A machine can generate a track. But only humans can bring lived experience, responsibility, and real relationships into its creation.

Developing an Artistic Voice Takes Time

You can generate a track in seconds.

Developing an artistic identity is different.

There is no fixed number of years after which someone suddenly becomes an artist. But finding your own voice usually requires a process of listening, learning, experimenting, completing work, and understanding which decisions actually belong to you.

You learn rhythm, melody, harmony, arrangement, performance, recording, sound design, mixing, and production.

More importantly, you develop taste.

You learn when to simplify.

You learn when to stop.

You recognize when something sounds impressive but does not feel honest.

You learn how to continue after rejection, disappointing numbers, creative blocks, technical problems, or periods in which no idea seems good enough.

You begin to understand what you want your music to communicate—and what you are willing to leave out.

Over time, hundreds of small decisions form something recognizable: your artistic voice.

As I explore in my article on how to find your signature sound as a music producer, that voice is not a preset, a plugin, or a production trick.

It is the result of repeated personal choices.

AI can imitate the surface characteristics of a style. It can recognize patterns and create something that resembles music that already exists.

But imitation of a result is not the same as living through the process that created it.

Music is rarely the work of one isolated person.

Music Is Also Collaboration

Behind many recordings is a network of musicians, singers, producers, songwriters, sound designers, recording engineers, mixing engineers, mastering engineers, visual artists, photographers, and label teams.

Each brings a different perspective.

Ideas change when another musician responds to them. A singer interprets a melody differently from the way it existed in the composer’s mind. A mixing engineer may reveal an emotional detail that was previously hidden. A mastering engineer listens from another perspective and helps the finished work translate beyond the studio.

Collaboration involves discussion, trust, disagreement, surprise, and compromise.

Sometimes the most meaningful moment in a track is something nobody planned.

That shared process is part of the art.

When an entire production is generated automatically, it does not only replace individual tasks. It can remove the conversation between creative people—the exchange through which everyone involved learns and the music becomes something no single person could have predicted.

Art is not only the result. It is also the process behind it and the relationships that grow through it.

If I Enjoy the Music, Why Should I Care Who Made It?

This is an understandable question.

If a piece of music makes someone relax, dance, concentrate, or feel something, does its origin really matter?

I do not want to divide listeners into those who make the “right” or “wrong” choices.

Music is personal. Everyone should be free to connect with whatever moves them.

My concern is that the choice should be informed—and that human artists should not become invisible beneath systems designed to maximize volume.

You can enjoy an AI-generated track and still want to know how it was made. You can appreciate an experiment while also believing that voices, identities, and copyrighted works should not be used without permission.

The emotional response of the listener is real, regardless of how the sound was produced.

But an emotional response does not answer the questions of authorship, consent, responsibility, or artistic identity.

When you support a human artist, you are not only supporting one audio file.

You are supporting the possibility of another album, a future collaboration, a concert, an independent studio, a creative career, and an artistic journey that can continue to develop.

You are supporting someone whose work may change because their life changes.

You can follow that person across years and understand how one release connects to another. You can read their story, see them perform, learn who contributed to the music, and enter a creative world with a real history behind it.

That relationship is one of the qualities that has always made music more than sound.

Listeners should at least be given the information needed to decide whether that relationship matters to them.

A Deezer-commissioned survey conducted by Ipsos among 9,000 people across eight countries illustrates this clearly. In a blind test using two fully AI-generated tracks and one human-made track, 97 percent of participants failed to identify all three correctly.

Yet 80 percent said fully AI-generated music should be clearly labeled, and 73 percent of streaming users wanted to know when a platform was recommending it.

You can examine the methodology and results in the Deezer and Ipsos study on attitudes toward AI-generated music.

People may not always hear the difference.

But many still care about knowing the difference.

A Polished Result Is Not Automatically Meaningful Art

AI-generated music can sound convincing.

That is one of the reasons this discussion is so difficult.

But the question cannot only be whether something sounds polished, pleasant, or technically correct.

Human-created music can also be generic or emotionally empty. AI-generated music may trigger a genuine emotional reaction in a listener.

The difference is not that one category automatically sounds good and the other automatically sounds bad.

The deeper difference concerns intention and authorship.

Who made the expressive decisions?

Who stands behind them?

Whose experiences informed the work?

Was anyone imitated without permission?

Was the listener given an honest description of what they were hearing?

The U.S. Copyright Office has made a related distinction from a legal perspective. Its report concluded that AI-generated output can be protected where a human author has determined sufficient expressive elements, but that merely providing prompts is not enough by itself.

You can read the summary from the U.S. Copyright Office on copyrightability and artificial intelligence.

But the distinction reflects a principle that matters creatively:

Requesting an outcome is not necessarily the same as authoring its expressive details.

When everything sounds polished, polish is no longer special. Identity becomes the scarce resource.

What Mass-Generated Music Risks Eroding

The danger is not that AI will replace every musician overnight.

The effects are more gradual.

It devalues creative effort

When music can be generated almost instantly, the years of learning, emotional work, financial investment, and personal development behind a human production become less visible.

The listener sees another three-minute track.

The process behind it is hidden.

It makes meaningful discovery harder

Every generated release enters an already crowded catalog.

Mass uploads can occupy search results, artist pages, recommendation systems, playlists, and databases. This makes it more difficult for listeners to find musicians who are attempting to develop sustainable careers.

The competition is not only for royalties.

It is also for attention.

It weakens trust

Listeners deserve to know whether an artist identity is real, whether a voice was authorized, and whether AI supported a human process or generated most of the work.

Without clear information, the relationship between the listener, the recording, and the supposed artist becomes uncertain.

It encourages sameness

Generative systems learn patterns from existing material.

They can produce music that is immediately familiar, technically convincing, and optimized for recognizable moods or functions.

But art often becomes meaningful through the opposite: unusual decisions, imperfection, cultural specificity, risk, contradiction, and ideas that initially appear difficult.

If platforms become filled with music designed to resemble what already works, we risk creating an endless reflection of the past.

It threatens collaboration

When entire productions are generated automatically, there may be no exchange between musicians, no interpretation by a performer, and no creative dialogue between producer and engineer.

The friction, humor, trust, and unexpected discoveries of working with other people disappear.

It separates creative value from responsibility

A human artist can explain their decisions, acknowledge influences, credit collaborators, respond to criticism, and stand behind a release.

An artificial artist identity cannot take responsibility.

Behind it may be a transparent experimental creator—or an anonymous operation producing content at scale.

The listener may never know which.

Consent, Credit, and Compensation

Another critical question remains:

What music was used to train the systems that generated the result?

Musicians, songwriters, publishers, record companies, and rights organizations have repeatedly called for transparency regarding training data, consent from rights holders, and fair compensation when copyrighted work contributes to a commercial AI system.

A global economic study commissioned by CISAC and conducted by PMP Strategy projected that, under unchanged market and regulatory conditions, as much as 24 percent of music creators’ revenues could be at risk by 2028.

This is a projection, not a guaranteed outcome. But it illustrates the potential economic transfer from creators to companies whose systems may depend on existing creative work.

The methodology and findings are available in the CISAC study on the economic impact of generative AI.

The principles should be straightforward:

Artists should be able to decide whether their recordings, compositions, performances, and voices may be used.

They should receive appropriate credit.

They should be compensated when their work contributes to commercial value.

Innovation should not depend on treating decades of human creativity as free raw material.

What Responsible AI in Music Should Look Like

The future does not have to be a choice between rejecting AI and accepting every possible use of it.

A responsible approach is possible.

It should include:

Clear labeling

Listeners should be able to distinguish between music that is substantially human-created, music that uses AI assistance, and music whose primary creative elements were generated by AI.

Consent

Artists, performers, and rights holders should be able to decide whether their work and voices may be used for AI training, imitation, or generation.

Credit and compensation

When human work contributes to the value of an AI system, the creators should not be excluded from the economic benefit.

Protection against impersonation

An artist’s voice, name, profile, photograph, and creative identity should not be replicated or exploited without authorization.

Action against fraud and spam

Streaming platforms and distributors must prevent mass uploads, deceptive profiles, manipulated streams, duplicates, and other attempts to extract money from the system unfairly.

Freedom for artist-led experimentation

Artists should still be able to explore AI creatively when they remain in control of the process and communicate its use honestly.

This is not anti-technology.

It is a framework for technology that respects the people whose work made its capabilities possible.

I Believe Human Artists Will Become More Valuable

Despite my concerns, I remain hopeful.

Human creativity will not disappear because a machine can generate sound.

The easier it becomes to generate unlimited music, the more valuable genuine identity may become.

When every platform is filled with technically acceptable tracks, technical acceptability is no longer special.

When music becomes infinite, meaning becomes rare.

What remains valuable is a recognizable artistic voice.

A human story.

An artist who evolves across albums.

A creative relationship that listeners can follow.

Music connected to real places, friendships, losses, discoveries, failures, and moments in time.

I believe many listeners will eventually grow tired of endless streams of interchangeable content.

They may begin asking different questions:

Who made this?

Why was it created?

Who performed it?

Who collaborated on it?

What does this music mean to the person behind it?

Can I trust that the identity and story are real?

In that future, “human-made” may become more than a description.

It may become a sign of trust.

The future of music should not be humans against technology. It should be technology that respects humans.

How You Can Support Human Artists

Listeners have more influence than they may realize.

Follow artists whose work matters to you. Save their albums and tracks. Share their music with other people. Read the credits. Subscribe to their newsletters. Attend concerts. Purchase music, vinyl, CDs, downloads, or merchandise when possible.

Supporting an artist does not only reward something that has already been made.

It helps make the next creation possible.

Artist-curated playlists can also offer a more direct and personal route into music than anonymous playlists filled with unfamiliar or unverifiable identities.

I personally curate my own Thomas Lemmer playlists, not only to present my work but to create thoughtful listening journeys through music I genuinely believe in.

Following playlists curated by artists you trust is one way to discover music with an identifiable human perspective behind it.

But the most valuable action does not have to involve my music.

Choose one human artist whose work means something to you.

Listen to one of their tracks today.

Then share it with someone who may connect with it too.

Small actions like these can help an artist continue.

Final Thought

AI can generate music.

It may generate something beautiful, useful, convincing, or emotionally effective.

But being an artist involves more than generating sound.

It means making choices, developing a voice, accepting responsibility, building relationships, and placing something of yourself into the work.

Music is communication between people.

That human connection is not old-fashioned.

It is the reason music matters.

And it is worth protecting.