By the end of September, Sony Music had asked digital platforms to take down more than 260,000 AI-generated tracks impersonating its artists.
That’s almost double the 135,000 tracks it had sought to remove by the end of March, according to the Financial Times, which reported the new figure today (October 5).
To slice the numbers another way: they suggest that, on average, Sony issued over 20,000 AI deepfake takedown requests per month in Q2 and Q3 (April 1 – Sept 30).
“We will keep enforcing our artists’ rights and expect platforms to act quickly and to prevent the same content from reappearing,” Sony told the FT. “But it’s becoming an uphill struggle.”
The FT did not report how many of the tracks platforms have removed, how quickly they acted, or which services received the requests.
Sony-affiliated artists targeted by the deepfakes include Harry Styles, Adele, Britney Spears, Queen, and Michael Jackson, the paper reported.
Sony described the problem to the FT as “generative AI deepfake tracks that imitate our artists’ voices and likenesses without permission,” which it said was “happening at scale, harms artists, misleads fans and undermines the ethical use of AI.”
In March 2026, the BBC described the tracks Sony had flagged over the previous year as songs “falsely purporting to feature” artists from its roster.
Spotify‘s impersonation rules separate two tactics: unauthorized AI voice clones and the fraudulent delivery of music, AI-generated or otherwise, to another artist’s profile.
Sony‘s published tallies do not distinguish between the two.
The proposed US NO FAKES Act describes a digital replica as “a newly created, computer-generated, highly realistic electronic representation that is readily identifiable as the voice or visual likeness of an individual.”
Dennis Kooker, President of Global Digital Business & US Sales at Sony Music Entertainment (SME), disclosed an early figure at an AI Insight Forum hosted by then-US Senate Majority Leader Chuck Schumer in November 2023.
- Kooker said then that SME had sent “close to 10,000 takedowns to a variety of platforms hosting unauthorized deepfakes that SME artists asked us to take down.” Kooker added that “platforms are quick to point to the loopholes in the law as an excuse to drag their feet or to not take the deepfakes down when requested.”
- By March 2025, Sony had requested more than 75,000 takedowns, according to its submission to the UK government’s consultation on AI and copyright law. That submission was obtained at the time by the FT and The Sunday Times.
- Then, in March 2026, Sony said it had asked streaming platforms to remove more than 135,000 deepfakes, in figures first reported by the BBC. Roughly 60,000 of those had been flagged since March 2025, according to the BBC.
“The problem with deepfakes are they are a demand-driven event,” Kooker said in March, according to the BBC. “They are taking advantage of the fact an artist is out there promoting their music.”
The FT‘s report did not include comment from streaming platforms.
In late 2025, Sony-owned distributor The Orchard issued takedown notices against I Run, a viral track by Haven, according to Billboard. The RIAA and IFPI also issued takedown notices against the track.
The track’s AI-processed vocals had drawn comparisons with Jorja Smith, whose label FAMM is distributed by The Orchard.
Haven’s Harrison Walker and co-producer Jacob Donaghue have said they used Suno to transform Walker’s own vocals, without specifically referencing Smith.
A Spotify spokesperson, cited by FAMM, said the platform “strictly prohibits artist impersonation” and that “this track was detected and removed, and no royalties were paid out for any streams generated.”
FAMM alleged that four different distributors had been used to deliver the track, in an attempt to sidestep the usual takedown process.
In September 2025, Spotify introduced an impersonation policy, stating: “Vocal impersonation is only allowed in music on Spotify when the impersonated artist has authorized the usage.”
Spotify also said it would cut wait times for reviews under its content mismatch process, which handles music wrongly delivered to artists’ profiles.
In March 2026, the same month as Sony’s disclosure, Spotify began piloting Artist Profile Protection, an opt-in feature that initially required artists to approve releases before they appeared on their profiles.
“Open-access distribution channels have lowered the barrier for independent artists to share music with the world, promote collaborations easily, and transfer music between distributors seamlessly,” Spotify said at the time. “But that openness comes with gaps that bad actors can exploit.”
Spotify has since revised the beta: in its standard setting, artists who opt in are notified when music is delivered under their name and can remove themselves from incorrect releases, which otherwise go live as scheduled.
Holding releases until an artist approves them is now an advanced setting with limited access, according to Spotify.
Fraudsters are also using AI to mass-produce tracks and then deploying bots to stream them, according to the FT. Those streams generate royalties that would otherwise flow to artists and rightsholders.
Kooker told Music Week in May that between 5% and 10% of streams could be fraudulent, though “no one knows exactly.”
The FT reported that label executives estimate streaming fraud could be costing the industry up to USD $2.2 billion annually.
Deezer said in July that it had received a daily average of nearly 90,000 fully AI-generated tracks in June – more than half of all new music delivered to the platform on peak days.
“AI streaming fraud is fast becoming an epidemic for the music industry, and the exponential growth should alarm everyone,” said Jamie Njoku-Goodwin, the former CEO of UK Music, in comments reported by the FT. “Whether it’s in deepfakes, streaming farms or AI slop being generated and distributed in bulk, organised fraudsters appear to be using platforms like Suno and DistroKid to rip off artists.”
The FT also cited a recent report by the Sonic Intelligence Academy, which found that 92.8% of AI tracks submitted to its registry had been made with Suno, and that DistroKid had been used to distribute more than three-quarters of those for which distributor data was available.
The report was based on 1,743 verified submissions.
In the US, cloning an artist’s voice does not clearly amount to copyright infringement, as MBW has previously reported.
The NO FAKES Act, which Sony supports, would create a federal right in a person’s voice and visual likeness.
To keep safe harbor protection under the NO FAKES Act, platforms would have to remove a flagged replica “as soon as is technologically and practically feasible.”
The bill cleared the Senate Judiciary Committee unanimously on June 18.Music Business Worldwide
