Introduction: A Silent Tipping Point
Throughout history, humanity has delegated its faculties: strength to the stone, time to the machine, memory to the digital. Today, it is delegating something deeper — creative thought itself. Generative artificial intelligence is no longer just a tool; it writes, composes, paints, arranges, sometimes with no genuine human intervention.
This is more than a technological shift. It’s a profound phenomenon of deindividualization. When humans no longer think, create, or even choose what they consume, but hand it all over to machines, the individual disappears — and with them, the very foundation of copyright: the presence of personality in a work.
- The Author: A Pillar of Law and Creativity
Copyright law is built on a simple principle: a work is protected if it is original, meaning it reflects the author’s personal imprint. This requires intention, choice, and human expression. It’s not just a technical issue, but a philosophical and legal one: a work without an identifiable author cannot be protected.
AI, however, produces texts, images, and music without intention, emotion, or unique style. It generates — it does not create. And that distinction is everything.
- The “Petit Papa Noël” Case: When Machines Usurp the Artist
In 2005, I represented Laurent Rossi, son of the legendary French singer Tino Rossi, in a landmark case. EMI, which had long distributed “Petit Papa Noël,” released a new version of the song — performed not by a real artist, but by a virtual character named Pinocchio, whose voice was generated via software, samples, and a poorly tuned recording of a child.
We argued that this “performance” was not a musical creation, but rather a synthetic construction, lacking personality, intention, or artistic quality. A Belgian court agreed, ruling that this version was not a genuine work and ordering its immediate withdrawal from the market. EMI later withdrew it in France as well.
-This case — years before the rise of generative AI — set a clear precedent: machines cannot substitute for humans in authorship, and the law must not be deceived by synthetic simulacra.
- AI: Built on the Works of Others
A fundamental issue with today’s AI systems is this: they are trained on massive databases of copyrighted works, without authorization or compensation.
It is undeniable that:
AI music generators rely on thousands of protected tracks to produce “new” compositions; Text-generating AIs like ChatGPT have been trained on millions of copyrighted books, articles, and scripts.
When Google indexed protected content, it had to pay for rights. When iPods, blank CDs, and hard drives enabled private copying, levies were applied. Why, then, should we tolerate AI systems exploiting authors’ works without compensation, all while generating commercial value?
- AI Does Not Create – It Plagiarizes in Fragments
Let us be clear: AI does not create in the legal sense. It reassembles fragments of existing works with minimal variation. If we were to ask an AI how it composed a musical piece, it might say:
“I combined Steely Dan’s harmonic base, a Daft Punk-inspired rhythm, and vocals treated like Billie Eilish.”
That is not creation — it’s composite plagiarism, an unauthorized use of protectable elements. This issue lies at the heart of recent lawsuits against OpenAI, accused of generating content “in the style of” well-known authors without their consent.
Even when users try to bypass this — asking, for example, for a “1970s pop song by a Swedish group called AFFA” — the AI will produce music clearly mimicking ABBA while formally denying it. This algorithmic mimicry is a legal façade, yet a factual reality.
- Market Saturation and Rights Bypass
The danger grows when we consider the sheer scale of AI output. Some platforms generate thousands of songs automatically — no effort, no talent, no limits.
One extreme case involved a company that copyrighted all possible melodic combinations using the 7 notes and 5 accidentals — a provocation showing that human creativity cannot compete with algorithmic brute force.
Worse, AI is now used to tailor songs for algorithmic promotion, targeting Spotify-like recommendation engines. This floods the market with artificial content, diluting real artists’ visibility and reducing their earnings.
- Spotify: Algorithmic Opacity and Playlist Hijacking
Spotify’s recommendation algorithms pose critical legal and ethical questions:
Free accounts cannot directly play a selected track. Users must first listen to another song, typically selected by the platform. Spotify modifies user-created playlists, inserting tracks by other — sometimes competing — artists. A curated “Best of Francis Goya” playlist may suddenly include unrelated titles, added by the algorithm.
These practices are:
unfair to users, damaging to artists, whose exposure is hijacked, and potentially legally questionable, as they raise the issue of whether such playlist manipulation violates moral rights or disrupts the editorial integrity of curated content.
- Transparency Obligations and the Spotify Scandal
Some digital service providers (DSPs) now require users to declare if AI was used in the creation of a track. This is a step toward transparency — but it is not always enforced.
The Spotify scandal revealed how thousands of AI-generated tracks, attributed to fake artists, were created and promoted. Some bore the same titles as famous songs, were objectively terrible, yet still amassed thousands of plays— artificially inserted into recommendation loops to reduce real royalty payouts.
This represents a systematic impoverishment strategy: flood the system, fake the plays, dilute legitimate income.
- Who Owns AI-Generated Works?
One core legal dilemma: who holds the rights to an AI-generated work? Three possibilities exist:
The AI itself? Legally impossible — AIs have no legal personality. The user who prompted the AI? Only if they made creative choices beyond just clicking. No protection at all? This is the current stance in many jurisdictions, including the U.S. and the EU.
This legal void creates a dangerous loophole: it encourages unregulated AI exploitation while preventing proper attribution and accountability.
The law must draw a firm line:
Only human-directed, intentional works deserve copyright protection.
- Can We Detect AI-Generated Works?
This is another growing challenge. There is no universal method yet, but several avenues are being explored:
Metadata analysis (when preserved), Stylistic fingerprints or repetitive patterns, Model-specific generation signatures, Development of AI detectors (e.g., by OpenAI or MIT).
Still, these tools remain fragile. The real risk lies in hybrid works — part human, part AI — blurring all boundaries.
Conclusion: Protecting the Human in the Age of Machines
We stand at a pivotal moment. If nothing is done, AI will create, select, and promote the music we hear, control the visibility of artists, and decide who gets paid. In such a world, authorship disappears, and with it, the meaning of creation.
Copyright is not just a financial mechanism. It is the legal recognition of human individuality. We must reinforce it — not weaken it — in the face of synthetic creation.
To defend authors is to defend humanity’s voice, and to ensure that, even in a world of machines, it is still the human trace that matters.
Cedric WAUCQUEZ
D.A.D. Dauntless Artists Defender