Skip to content
FREN
Menu

News · Impact studies

What studies tell us about AI's impact on music creators

Since 2024, several studies have tried to estimate what generative AI could cost creators in income and jobs, and one of them identifies library music as the most exposed. We set out the main projections, who commissioned them and their limitations.

How to read these figures

None of the figures that follow measures a loss already incurred. They come from projections and surveys, most of which were commissioned by organisations that represent creators' interests. This does not make them wrong, but it is worth stating. We report the figures as published, just as we do with platform figures, and give the commissioning organisation, publication date and methodology for each study.

CISAC estimates a quarter of music revenues could be at risk by 2028

CISAC (the International Confederation of Societies of Authors and Composers) commissioned this study from the consultancy PMP Strategy. Published on 04/12/2024, it is presented as the first global study to measure the economic impact of generative AI on music and audiovisual production (CISAC, 04/12/2024).

The study projects that 24% of music creators' revenues and 21% of audiovisual creators' revenues could be at risk by 2028. Cumulative losses would reach €22 billion over five years, with €10 billion in music and €12 billion in audiovisual production. By then, AI-generated music would account for about 20% of streaming platforms' revenues and about 60% of revenues from library music catalogues. Library music is licensed for use in documentaries, news reports, advertisements and many other audiovisual productions.

In this projection, library music is therefore the segment most at risk. It is also where losses would be most likely to go unnoticed, since nobody reads an advertisement's credits to find out who composed the music.

€2.7 billion at stake in France and Germany, according to Goldmedia

Goldmedia conducted this study for Sacem (France) and GEMA (Germany). Published on 30/01/2024, it combines market analysis, expert interviews and a survey of more than 15,000 music creators and publishers (CISAC, 30/01/2024).

It estimates that 27% of music creators' revenues would be at risk by 2028 if no system were introduced to pay for works used by AI. For the members of the two societies, cumulative losses would reach €2.7 billion by 2028, including €950 million in 2028 alone.

Of the creators surveyed, 71% believe that AI threatens their ability to make a living from their work, and 64% consider that the risks outweigh the opportunities. AI is already used by 35% of respondents, rising to 51% among those under 35, and 95% call for more transparency from AI developers.

UNESCO cites similar projections for global losses

UNESCO's global report Re|Shaping Policies for Creativity, published on 18/02/2026 and covering more than 120 countries, cites projected revenue losses of 24% for music creators and 21% for audiovisual creators by 2028. Among the cultural professionals consulted for MONDIACULT, 79% see AI as a threat to people working in the arts (UNESCO, 18/02/2026).

CVL Economics' projections for US entertainment jobs

The study Future Unscripted, conducted by CVL Economics for the Animation Guild, the US animation union, was published in January 2024. It is based on a survey conducted in late 2023 among 300 executives across six US entertainment industries (The Decoder, 01/2024).

It projects that around 204,000 jobs in US entertainment would be disrupted by generative AI by 2026, including 118,500 in film, television and animation (21% of the jobs in those sectors) and about 1,800 in music and sound recording. A disrupted job is not necessarily a lost one, since it may be transformed, merged with another or scaled back. The study also measures executives' intentions rather than job losses actually observed.

Early evidence from the labour market

The evidence available so far presents a more nuanced picture than the projections. A study published in the Journal of Cultural Economics, based on US labour market data from 2017 to 2024, finds "little evidence so far that generative AI has broadly reduced artists' earnings". It nevertheless identifies composers and music directors as among the artistic occupations most exposed to generative AI (Gallup, 03/05/2026).

The losses projected for 2028 have yet to show up in the aggregate statistics. This is why artists, record labels and publishers need transparency and verifiable data now, rather than having to wait for a retrospective assessment. The observatory of disputed decisions is our contribution to that data.

Sources

Updated 06/10/2026