Where does the human remain when artificial intelligence becomes part of the creative process?
This question shaped a full-day online conference that brought together musicians, vocalists, visual artists, photographers, and filmmakers—people from different disciplines, with different experiences and very different attitudes toward artificial intelligence.
In recent months, more platforms and distributors have begun asking creators to disclose whether AI tools were used in a work. That could have been an important step toward transparency. But the system remains too crude. Instead of showing what a person created and where a tool assisted them, it often reduces an entire creative process to two buttons: yes or no.
Two buttons cannot describe a creative process
How can an author select “no” if AI was genuinely used—perhaps in sound processing, translation, layout, or the preparation of a visual element?
And how can that same author select “yes” when the label leads people to assume that artificial intelligence created everything and that no meaningful human work remains?
A binary disclosure may identify the presence of a tool, but it does not identify authorship. It does not tell us who wrote the words, composed the melody, performed the parts, created the source material, made the artistic decisions, or remained responsible for the final work. The technology becomes visible. The author disappears.
At what point did Masha stop being the author?
Imagine that Masha kept personal diaries for several years. They contained her life: events, thoughts, feelings, memories, and observations.
One day, she decided to publish them. She could not afford a large publishing team, so she used accessible AI tools. With their help, she divided the text into chapters, developed subheadings, chose fonts, prepared the layout and table of contents, made a cover and several illustrations, and perhaps translated the book into five languages.
Who is the author of that book? At what precise moment did Masha lose her authorship? When she divided the text into chapters? When she chose the typeface? When she created the cover? When she translated the book? Or when she honestly disclosed that she had used artificial intelligence?
AI did not live Masha’s life. It did not keep her diaries for years. It did not experience her love, loss, fear, or hope. It helped her prepare her work for publication; it did not turn her into a stranger standing beside her own book.
The same thing is now happening in music, literature, photography, and visual art. A person creates the work and makes hundreds of creative decisions, then uses AI at one stage of the process. A single label can make the audience imagine that the person merely typed three words and received a finished work.
If a system cannot say when Masha stopped being the author, it should not erase her authorship with a single label.
Transparency or punishment for honesty?
We are not opposed to disclosure. On the contrary, meaningful transparency is necessary.
But transparency should reveal the actual distribution of creative and technical contributions. It should show who wrote the text, composed the music, performed the parts, created the original materials, used particular tools, made the decisions, and accepted responsibility for the result.
A general label saying “created using AI” explains none of this. Worse, the current system can make honesty a disadvantage. A creator who voluntarily discloses the use of AI may immediately face suspicion, diminished value, exclusion from certain categories, or reduced visibility. A creator who hides the same use can remain safely classified as human-made.
The label then measures not how a work was created, but how candid its author was.
If honest disclosure turns a creator into an outcast, it does not create transparency. It teaches people to hide the truth.
What is disclosure actually for?
Disclosure should not divide art into what is supposedly pure and impure. It should help us understand what is happening.
Which tools are professional creators using? At what stages do they provide genuine assistance? Where does human authorship remain? Which practices raise legal or ethical problems? Where do source materials, performers, voices, and artistic identities require protection? How is creative practice itself changing?
Reliable information is essential. Automated detectors may provide signals, but they cannot reconstruct an entire creative process or measure every human contribution with certainty. If honest creators become afraid to disclose their methods, the resulting statistics are fundamentally distorted.
We cannot design intelligent regulation from data that the system itself encourages people to conceal.
Disclosure should be an instrument of knowledge, not an instrument of punishment.
Human-made does not automatically mean meaningful
The conference repeatedly returned to the idea of protecting human-made work. But what, exactly, are we protecting?
The words human-made guarantee neither talent nor professionalism nor artistic value. A work made entirely by a person may still be crude, derivative, or empty. Conversely, the use of a contemporary tool does not destroy human intention, experience, judgment, or talent.
It is possible to type a three-word prompt and receive a random composition. That does not mean every musician working with AI creates music in that way. Artists, photographers, writers, and musicians can use the same technology in radically different ways. For one person it is a button for rapid content production. For another it is one instrument inside a demanding professional practice.
The real dividing line is not between a human being and AI. It is between intentional authorship and thoughtless content production.
Protecting human creativity cannot mean forbidding people to use contemporary tools. Otherwise, we begin defending an imaginary sterile process instead of defending the author.
Who actually pays for this fight?
Disclosure policies and restrictions are often presented as a way to stop industrially generated content, streaming farms, and businesses built on artificial plays. Those problems are real.
Content can be generated and uploaded at an industrial scale. If one track is removed, dozens more may already have appeared. If one account is blocked, another can be created. For such a business, a song is an expendable unit and enforcement risk is simply part of the model.
Yet the consequences of broad restrictions are often felt most sharply elsewhere—by the least protected people in the system: independent artists.
An independent musician with several thousand monthly listeners usually earns nowhere near enough from streaming to cover production costs. Professional software, subscriptions, equipment, distribution, artwork, and promotion all have to be paid for. Even modest promotion for one song may cost hundreds of euros while streaming income amounts to only a few dozen dollars.
For a content farm, a deleted track is one row in a spreadsheet. For an independent artist, it may represent months of life, work, and personal investment.
And it is this artist who may be penalized for being transparent: excluded from recommendations, removed from algorithmic opportunities, treated as less authentic, or made to carry a label that conceals rather than explains their contribution.
Industrial systems will adapt. An independent author may not survive one more restriction.
Measures designed to fight the industrial production of musical waste can end up striking hardest at those for whom every song is a real work.
The question we could not answer
The conference included supporters of AI, opponents of AI, and people whose position lay somewhere between. Yet, slowly, we arrived at one shared conclusion: the human being must remain at the centre of every decision.
Not the technology. Not the label. Not a formal yes-or-no box. The person, their source materials, their creative decisions, their work, and their rights.
We can support independent artists. We can demand more precise acknowledgement of human contributions. We can speak openly about creative processes and refuse to let an author be erased from their own work.
But individual artists cannot stop industrial generation, streaming farms, or the economy of artificial plays. Those questions belong to legislation, distributors, streaming services, corporate financial interests, and international regulation.
One question therefore remained open: how can independent creators protect themselves within a system whose rules are shaped by institutions with an economic interest in preserving its current structure?
We do not have a complete answer today. But we do know this: human creativity cannot be protected by destroying the least protected human creators.
Transparency must keep the human being visible within the work, not erase them from it. If we genuinely want to defend human-made art, we must begin by defending the human being.