Music generated by AI could put 24% of music creators’ revenues at risk by 2028, resulting in an annual loss of earnings of €4 billion. Meanwhile, the market value of AI-generated music in 2028 could top €16 billion, with the music AI platforms also bringing in about €4 billion in fees and subscriptions.
This is all according to a new study by CISAC, the global grouping for songwriter collecting societies. Because it’s a CISAC study, ‘music creators’ means songwriters. Additional losses would be made by artists on the recordings side.
The losses incurred by the music creator community at large could be mitigated, to an extent, if artists and songwriters also shared in the revenue streams of AI music.
That music is usually generated by AI models that were trained by being exposed to existing music. If the AI companies had to get permission from the music industry to use existing music in their training, that would provide an opportunity for creators and rightsholders to demand a share of future income.
As far as the music industry is concerned, when copyright protected works are used, that permission is already required, but many AI companies disagree. To that end, the music industry wants lawmakers to clarify the copyright obligations of AI companies and this report from CISAC will be used to argue that that clarity is now needed urgently
Indeed, responding to the report, Roberto Neri, CEO of The Ivors Academy, the organisation representing UK songwriters, says, it’s now “urgent that legislators act”.
“Songwriters and media composers face losing nearly a quarter of their income by 2028 due to generative AI", he adds, "while AI companies stand to gain €4 billion from the unlicensed reproduction of creators’ works. Transparency, fair pay and control are non-negotiable”.
“Creators must be at the centre of legislative decisions shaping the AI landscape”, he goes on. “We urgently need rules that ensure transparency, fair remuneration, consent and control”.
The CISAC report was produced by PMP Strategy and informed by interviews and workshops with over 50 industry professionals.
In estimating the market value of AI-generated music in 2028, it was assumed that this music will account for a portion of consumption on streaming platforms - especially with mood music and functional audio - and a significant portion of library music used by media and businesses producing audio-visual content.
The report predicts a steep increase in that market value over the next few years, from €1 billion last year to €16 billion by 2028, generating a total of €40 billion over the next five years.
In terms of the negative impact on music creator earnings, again this will rise over the next five years, totalling around €10 billion over that period, and €4 billion in 2028.
This is based on creators earning less as AI-generated music accounts for increasing amounts of digital consumption and is increasingly used by media and businesses producing audio-visual content.
Many generative AI companies have trained their models on unlicensed content, arguing that they do not need permission from copyright owners because of data mining exceptions or fair use principles in certain copyright systems. The music and wider copyright industries are adamant those exceptions do not apply, and that AI companies must negotiate licensing deals with the copyright owners.
If - as a result of copyright owners winning test cases in court or getting clarity from lawmakers - the AI companies do start to negotiate such deals, the next question is what those deals look like and whether creators and rightsholders will get a one-time training fee or will be cut into future income.
If the latter - and it seems likely that’s what the copyright industries will push for - they could try to negotiate a share of the market value of AI-generated content and/or the revenues directly generated by the AI platforms themselves.
The CISAC report doesn’t get into the legalities or deal making, but the figures will clearly be used by the music industry as it seeks the support of governments and legislators to help force the AI companies to the negotiating table.