Get all your news in one place.
100’s of premium titles.
One app.
Start reading
MusicRadar
MusicRadar
Entertainment
Andy Price

Sony Music says it has removed over 75,000 deepfake tracks from streaming platforms, which include voice models of Beyoncé, Harry Styles and Queen

Harry Styles.

As one of the music industry’s ‘big three’ record labels, Sony Music is understandably on edge when it comes to the softening of the rules and regulations around AI training. As part of a recent consultation being carried out by the UK government, Sony shared that it had isolated and moved to take down over 75,000 examples of tracks that purport to feature artists on their roster, but are, in fact, AI-generated.

The duplicated artists include the likes of Beyoncé, Harry Styles and Queen - all big financial pillars of the label. The artists' voices were algorithmically modelled and used to sing/perform original or AI-generated songs, which were then uploaded to steaming platforms.

As Sony told The Financial Times, the quantity of songs generated in this way is only increasing as technology (and access to deepfake technology) gets cheaper and easier.

Sony states that the 75,000 taken down so far probably only equates to a small percentage of the overall number of deepfaked tracks currently out there.

With protection of their artists' copyright of paramount concern, the label has been forced to manually browse and find them, and issue takedown requests to the uploader directly. A major time-sink.

Queen are among the Sony Music artists that have been heavily deepfaked (Image credit: FG/Bauer-Griffin/Getty Images)

Sony's statement and data sits with a developing political context.

Currently, the UK government is keen to loosen regulations around AI, in the hopes that it will encourage new businesses and start-ups to be more flexible with this emerging technology. In the words of UK Prime Minister Keir Starmer, the aim is to make Britain ‘the best place to start and scale an AI business’.

The government is currently mooting a dynamic which would give the green light for AI companies to train their models (for no cost) on a huge array of copyrighted music, film, books and newspapers. The onus being on the copyright holder to make a point of contacting these companies and opting-out their work.

Protest to this proposal has been loud - and in the most popular response, extremely quiet.

While the government has yet to come to a final decision on how it deals with copyright infringement in this new context, indications are that they’re gravitating increasingly towards open-access models.

The FT has taken a look at Sony’s submission to the UK government’s consultation, within which it states that AI-generated recordings of label IP result in “direct commercial harm to legitimate recording artists, including UK artists.”

Sony also stated that the government’s leaning towards an open-door policy was “rushed, unbalanced and irreversible.” It made the assertion that copyright was “a right, not a regulation,” and a “necessary societal reward” for creating music.

Sony fear that the proposed erosion of UK copyright law will not just affect bigger, commercially successful acts, but will have an impact on smaller artists who are sustained by streaming revenue, and don’t have a label like themselves backing them up.

But will the outcry from one of the industry’s most significant stakeholders be enough to give pause to the government’s open-access sympathies?

Sign up to read this article
Read news from 100’s of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
One subscription that gives you access to news from hundreds of sites
Already a member? Sign in here
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.