Skip to content
FREN

An open and independent platform devoted to the human provenance of music

The right to know where music comes from.

Every day, tens of thousands of AI-generated tracks reach streaming platforms. At the June 2026 peak, fully AI-generated tracks made up more than half of everything delivered to Deezer. Without campaigning against AI, we stand up for human artists and for everyone's right to know where the music they hear comes from.

WORKmusic · lyricsCompositionRecordingvoice · productionCreationCirculationListeningAppeal

Key figures

90,000

fully AI-generated tracks a day were delivered to Deezer at the June 2026 peak.

Source: Deezer, 21/07/2026

24%projection

of music creators' revenues could be at risk by 2028.

Source: CISAC and PMP Strategy, 04/12/2024

no

appeal procedure is common to all streaming platforms and distributors.

Source: VERIMUSIC review, 29/09/2026

From creation to listening

Creation

At the centre is the work, born of two processes that do not always have the same origin.

Compositionmusic, lyrics

What gets written: the melody, the harmony, the words. By whom, and with which tools.

Recordingperformance, production

What you hear: a voice, a performance, a mix. A voice can be cloned.

Three possible origins for each process

  • Human creation
  • AI-assisted
  • AI-generated

A generated composition can be sung by a human; a cloned voice can carry a song written by a human.

Our demands

  • Recognition of hybrid creation
  • Consent to the use of voices and works

Circulation

Distributors and platforms receive the work, detect it, label it, sometimes block it, and distribute the revenue it earns. Their rules vary from one service to the next, and none of their detectors has, to our knowledge, been independently audited.

Our demands

  • Detectors open to independent scrutiny
  • Revenue for human artists

Listening

Every listener must be able to see at a glance whether and how AI was used to make the track they are listening to. Article 50 of the EU AI Act, which concerns transparency for AI-generated content, has applied since 02/08/2026.

Our demands

  • The right to know

An effective right of appeal

If one of your tracks has been labelled, blocked or demonetised and you dispute the decision, tell us what happened. We will only publish your name with your consent.

Observatory of disputed decisions

The manifesto

The right to know where music comes from

“We stand up for human artists and the music they create, and for the people who listen to it, who have every right to know where it comes from.”
Adel Hendaoui, PhD in Information Systems, founder of VERIMUSIC
  1. 01

    The right to know.

    Every listener must be able to see at a glance whether and how AI was used to make the track they are listening to.

  2. 02

    Detectors open to independent scrutiny.

    Those who detect AI-generated music must publish their methods and error rates and have them checked by independent third parties.

  3. 03

    An effective right of appeal.

    Artists must be able to challenge any decision that blocks, labels or demonetises a track. They must be told why the decision was made, have a real person they can contact, and know what evidence will be accepted and when to expect a response.

Read the full manifesto and its six demands

Sign the manifesto

·

signatures so far

Only signatures confirmed by email are counted.

Add your name

We use your email address only to confirm your signature. It is never published, and we will not send you any newsletter unless you ask for one. Full details are on our Privacy page.

Sections

  • Take part
  • Stay informed
  • Resources
  • Project
  1. The method and protocol we are developing to document the provenance of a piece of music will be published under an open licence for anyone to use.Resources · 01Open tools
  2. If one of your tracks has been labelled, blocked or demonetised and you dispute the decision, tell us what happened. We will only publish your name with your consent.Take part · 02Observatory of disputed decisions
  3. How streaming services and distributors detect, label or block AI-generated music, and what options artists have for appealing their decisions.Resources · 03Platform rules
  4. Who to contact in your situation, what evidence to gather before writing, and a template letter to challenge the decision.Resources · 04Artist's guide
  5. What these standards and rules cover, and when they apply. Article 50 of the EU AI Act, which concerns transparency for AI-generated content, has applied since 02/08/2026.Stay informed · 05Standards and regulation: C2PA, DDEX, AI Act
  6. The detection tools in use, the accuracy their makers claim for them, and the absence, so far, of any known independent audit.Stay informed · 06Detectors
  7. The vocabulary of the debate, from human provenance to voice cloning, explained simply.Resources · 07Glossary
  8. Who is behind the project, when it began, and why it is independent of platforms.Project · 08About
  9. Decisions by platforms, distributors and lawmakers over recent months, in the order they were made.Stay informed · 09News and analysis

Published projections

These are estimates for 2028, not losses that have already been recorded.

60%

of revenue from library music used in documentaries, advertising and news reports could go to AI-generated music in 2028.

Source: CISAC and PMP Strategy, 04/12/2024

€2.7bn

in cumulative losses by 2028 for music creators and publishers in France and Germany.

Source: Goldmedia for Sacem and GEMA, 30/01/2024
The studies and their limitations