Sony removes a staggering 135,000 deepfake songs from music streaming services, but ‘labelling AI material is absolutely the next critical challenge’

• Sony Music pulls over 135,000 deepfakes impersonating its biggest artists
• This includes acts like Beyoncé, Harry Styles, and Bad Bunny
• Now industry giants are calling on music platforms to clearly label AI-generated content

The battle between the best music streaming services and AI-generated slop is still an ongoing conflict, and now Sony Music is getting involved — it’s removed over 135,000 deepfake songs of some of its most famous artists.

According to the music giant, the generative AI deepfakes were created to impersonate a slew of its biggest artists, including Harry Styles, Beyoncé, and Queen, with other artists such as Miley Cyrus and Bad Bunny potentially being affected. This figure was disclosed at the launch of the Global Music Report by the International Federation of the Phonographic Industry (IFPI) earlier this week.

In a statement issued to the BBC, president of Sony's global digital business Dennis Kooker detailed the harm these deepfakes can cause, sharing, "In the worst cases, (the deepfakes) potentially damage a release campaign or tarnish the reputation of an artist”.

Artist reputation aside, fraudulent AI slop causes huge damage to an artist’s financial gain. Kooker added, "That is when deepfakes are at their worst - building off and benefiting from the demand the artist has created (and) ultimately detracting from what the artist is trying to accomplish.” But the figures also reveal how fast AI slop and fraudulent streams are increasing.

Though Sony has removed the 135,000 AI-identified tracks, it’s possible this is only a portion of AI-generated content being uploaded across the likes of Spotify and Apple Music. Apparently, Sony Music flagged around 60,000 songs falsely claiming to be from its pool of artists since March last year, which the company also revealed in the report.

As it stands, music fans are running into AI-generated content on streaming platforms more often than they should be, which can be attributed to the fast development of AI models and their inexpensive accessibility. We’ve previously reported on this common issue with Spotify subscribers — but it’s not the only platform where AI slop is building up.

It’s got to the point where music industry giants are now actively calling on the leading streaming platforms to identify and flag AI-generated content, but while there are some measures already in place, it’s still a bit of a gray area.

Labeling AI content is ‘the next step’

In the BBC’s report, the company shares a quote from the CEO of the IFPI, Victoria Oakley, who says, "I hate to say it, but it's very simple to fix”, calling for music platforms to be more transparent with labelling content that’s AI-generated. As a music buff, I can say on behalf of listeners everywhere that this is what we want, but even the biggest platforms still haven’t wrapped their heads around how to efficiently flag AI music.

Deezer was the first streaming giant to roll out its own AI-detection tool, which explicitly labels songs that are 100% AI-generated, excluding them from your playlists — but the catch is that Deezer doesn’t pull them from the service altogether.

Recently, Deezer shared its business earnings report, where it revealed that 85% of AI-generated fraudulent streams were demonetized thanks to its proprietary detection technology. Similarly to Spotify, which actively encourages users to report AI slop, Deezer acknowledges that there’s a distinction between songs that are 100% AI-generated and songs where only a fraction has been generated using AI for creative purposes, but this opens a dangerous loophole.

Prior to uploading a song to Deezer, if a creator uses AI to generate an entire song but then makes manual production changes and enhancements to it themselves, Deezer will not flag that song as being AI-generated. So, it’s a bit of a tricky one.

That said, Deezer’s system could be seen as the more reliable one compared to Apple Music, for example, which is rolling out its own ‘Transparency Tags’. The biggest difference with Transparency Tags is that Apple Music is leaving the responsibility of disclosing AI content to labels and distributors, an option which they can easily reject, while Deezer's system doesn't give you a choice.

So while Deezer's system does raise some slight flaws, it still has a one-up on Apple Music's approach, and also Spotify, which is yet to roll out AI content tags.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course, you can also follow TechRadar on YouTube and TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.