Sony Music has sought the removal of more than 260,000 AI deepfake tracks from digital platforms, a total that highlights the rapidly expanding challenge of unauthorized artist impersonation.
The figure is nearly twice the number disclosed in March, when Sony said it had submitted 135,000 takedown requests. The company says the targeted recordings use generative AI to copy its artists without their consent.
A Sony Music spokesperson said the tracks “imitate our artists’ voices and likenesses without permission,” describing the volume of material as almost double what the company encountered earlier in the year.
“It’s happening at scale, and harms artists, misleads fans and undermines the ethical use of AI,” the spokesperson said. “Every fraudulent stream takes money away from the artists and songwriters whose music is being listened to.”
Sony said it will continue enforcing its artists’ rights and expects digital platforms both to respond quickly and to prevent removed material from returning. However, the spokesperson called that effort “an uphill struggle” and said the company is working with industry partners to improve the detection and removal of unauthorized deepfakes.
A patchwork of protections
The mounting number of takedown notices comes as the United States remains without a federal law specifically designed to prevent or stop the unauthorized use of an artist’s name, image, likeness or voice in an AI deepfake or soundalike recording.
Protections currently vary at the state level, creating an uneven system that can depend in part on where the deepfake originates. Against that backdrop, major music companies and streaming services began working together in 2023 on a notice-and-takedown process.
That voluntary procedure allows trusted labels to request the removal of deepfakes involving their artists. It is a stopgap rather than a nationwide legal standard, but the system has helped companies including Sony respond to the growing volume of impersonation tracks appearing on digital services.
Sony’s latest numbers also show the scale of the policing burden falling on labels. Beyond requesting initial removals, the company wants platforms to stop the same recordings from resurfacing after they have been taken down.
The push for the NO FAKES Act
The music industry is also seeking nationwide protections through the NO FAKES Act. The proposed legislation would give individuals the right to authorize the use of their voices and likenesses in AI deepfakes. It also includes exclusions covering areas such as news, parody, documentaries and sports.
NO FAKES was initially introduced as draft legislation in 2023 before being formally introduced for the first time in 2024. The music business has continued pushing for its passage since then.
The bill was brought to the Senate floor last week and received nearly unanimous support. Senator Ted Cruz of Texas blocked it, however, citing concerns about the use of a person’s image in satire or political commentary.
That legislative debate is unfolding while digital platforms face a huge daily influx of AI-created music. French streaming service Deezer says more than 90,000 fully AI-generated tracks are uploaded to its platform every day.
Only a fraction of those uploads are considered AI deepfakes. A fully AI-generated recording does not necessarily impersonate an existing artist, while the material targeted by Sony allegedly copies artists’ voices or likenesses without permission. Even so, the volume of AI music illustrates the detection challenge as labels try to identify the smaller category of tracks involving artistic impersonation.
What Happens Next?
Sony says it will keep pursuing unauthorized deepfake recordings while collaborating with partners across the music industry to detect and stop them. Digital platforms will remain central to that effort through the existing voluntary takedown system.
At the same time, the industry’s campaign for the NO FAKES Act continues after the bill’s latest Senate setback. Until a nationwide standard is established, labels must keep relying on the current mix of state protections, platform cooperation and their own enforcement efforts as AI-generated uploads continue to multiply.