Sony Music Entertainment has taken down more than 100,000 AI-generated deepfake songs falsely attributed to its artistes.

The development follows a report from the International Federation of the Phonographic Industry, which in its latest global music report said fraudulent streaming practices are diverting revenue away from legitimate artistes.

The report noted that artificially generated songs used for manipulated or fake content are “siphoning vital revenues” away from the actual artistes.

According to the BBC, Sony said the spread of deepfake music is causing “direct commercial harm to legitimate recording artists”, particularly targeting musicians during their album promotion stage.

Dennis Kooker, president of Sony’s global digital business, said, “In the worst cases, the deepfakes potentially damage a release campaign or tarnish the reputation of an artist.”

He added that fake tracks often surface mostly when an artiste is actively promoting new music.

“It’s taking advantage of the fact an artist is out there promoting their music. 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,” Mr Kooker said.

Sony disclosed it had identified about 135,000 deepfake tracks so far, with roughly 60,000 discovered since March 2025 alone. The company believed the figure showed only a fraction of the total fake content uploaded across streaming platforms.

Artistes such as Bad Bunny, Miley Cyrus, and Mark Ronson are among those potentially affected.

The IFPI said artificial intelligence has “supercharged” streaming fraud, making it easier and cheaper to produce fake songs that can convince and mislead fans, thereby diverting the artiste earnings.

Victoria Oakley, CEO of the IFPI, urged the music community to take action across the industry, describing streaming fraud as “theft, plain and simple”.

“The entire music community must take action to tackle the threats facing our industry. The organisations with the data, scale and leverage to prevent this fraudulent activity, including streaming services, content aggregators and distributors, must take decisive action,” Ms Oakley said.

Mr Kooker pointed to efforts by music streaming platform Deezer, which has introduced tools to identify AI-generated tracks.

The platform estimates that about 34 per cent of submitted songs are now categorised as AI-generated.

Mr Kooker said without proper identification, listeners may struggle to distinguish between genuine human creativity and AI-generated content.

“Transparency shouldn’t be optional. It’s the foundation of a fair and sustainable music ecosystem,” he said.