Khaberni - The battle between the best music streaming services and AI-generated content continues, as "Sony Music" has stepped in by removing over 135,000 fake songs using deepfake technology for some of its most famous artists.
Major artists are victims of deepfake
According to the music giant, these clips were created using generative artificial intelligence to impersonate the voices and works of many of its top stars, including Harry Styles, Beyoncé, and the band Queen, with other artists like Miley Cyrus and Bad Bunny potentially affected. This figure was revealed during the launch of the International Federation of the Phonographic Industry (IFPI)'s global music report earlier this week.
Damage extends beyond reputation to financial losses
In a statement to the "BBC", Denis Koker, President of Global Digital Business at Sony, explained the extent of damage that these clips could cause, saying: "In the worst cases, these fake works could negatively impact the launch campaigns of artistic works or harm the artist's reputation."
And it's not just about reputation, as these fraudulent materials cause substantial financial losses for artists.
Numbers reveal the acceleration of the phenomenon
Although Sony removed 135,000 clips identified as AI-generated works, this number is likely just a fraction of the content being uploaded to platforms like Spotify and Apple Music.
The company also noted that it had detected about 60,000 songs falsely claimed to be from its artists since March last year, according to the report.
Currently, music lovers are increasingly encountering this type of content on streaming platforms, attributed to the rapid development of artificial intelligence models and their low-cost accessibility. This issue has been highlighted among Spotify users, but it is not limited to a single platform.
Calls for clear labeling of AI-generated content
The matter has reached the extent that major players in the music industry are demanding streaming platforms to clearly identify and distinguish AI-generated content. Although some measures are already in place, the issue remains largely in a gray area.
In the "BBC" report, Victoria Okley, the CEO of the International Federation of the Phonographic Industry, was quoted as saying: "I hate to say this, but the solution is very simple," urging platforms for more transparency in labeling AI-generated content. From the listeners' perspective, this is a fundamental demand, but even the largest platforms have yet to find an effective mechanism for its implementation.
Deezer takes a step ahead in detecting content
Deezer was the first major streaming platform to launch a special tool for detecting AI-generated content, where it places a clear mark on songs made entirely by artificial intelligence and excludes them from playlists, but does not delete them from the platform entirely.
Deezer revealed in its latest earnings report that 85% of fraudulent streaming operations linked to artificial intelligence had their revenue generation cancelled thanks to its detection technology. Like Spotify, which encourages users to report this type of content, Deezer acknowledges a distinction between songs produced entirely by artificial intelligence and those where artificial intelligence is partially used for creative purposes, which creates a dangerous loophole.
Loopholes in classification
For example, if a content creator produces a song entirely using artificial intelligence and then makes manual adjustments to it later, it might not be classified as an AI-generated work on the Deezer platform, complicating the process of distinction, according to "Tech Radar".
Despite some gaps, Deezer's system is still considered more reliable compared to Apple Music's approach, which has begun introducing what are known as "transparency tags". The main difference is that Apple Music leaves the responsibility of disclosing artificial intelligence use to production companies and distributors, an option they can easily ignore, whereas Deezer does not allow this option.
While Deezer's system is not without some loopholes, it remains more effective than the approaches of Apple Music and Spotify, which have not yet launched a clear system for labeling AI-generated content.
