On a late February afternoon, singer Ankur Tewari excitedly opened Spotify for Artists, a dashboard that allows musicians to view their content on the streaming app, to check how a new song had been received. Chand Takiye, the single, was still fresh and the latest of the soft, reflective tracks that had earned Tewari a fanbase.
Instead, Tewari was stunned to find a set of songs he had never written, recorded or approved sitting under his name on Spotify, and also on Apple Music.
Tewari is not an artiste easy to mistake for someone else. Over the years, his songs, such as Dil Beparvah and Gehraiyaan, have found a distinct consumer base in the indie pop and acoustic genres. Now, these dozen or so random tracks, composed by an impersonator copying his style and wearing his identity, were garnering playtime, listing him as the main performer.
“The tracks hijacked my audience reach but credited a certain Henry Hobson as composer,” Tewari told Mint. “This let the impersonator collect the publishing royalties, while my artist page absorbed the poor performance signals from the fake streams.”
Tewari says the tracks weren’t just misattributed but appeared to be generated and uploaded at scale. His own search suggested a likely pathway where songs are generated instantly by artificial intelligence (AI) tools such as Suno, and then uploaded through distributors by tagging an existing artiste’s name.
Tewari and his team moved quickly to get the fakes taken down, but it took all of three days. And in streaming, three days can be long enough to matter.
“Tracks that perform poorly can distort the data that platforms use to read an artiste’s engagement and momentum and those signals do not always disappear even when the songs do,” says Tewari.
For many artistes, this isn’t an isolated shock but a recurring phenomenon now amplified by AI generated music.
Varun Rajput, frontman of the rock band Antariksh, says misattributed tracks have appeared on his profiles way too many times, sometimes as random devotional or regional songs sitting awkwardly alongside a rock catalogue.
“It’s pretty annoying, especially when fans point it out,” he said, adding that getting such tracks removed often involves manual forms and outreach to platform teams.
These incidents point to a new and murky vulnerability in the digital music space. A fake artiste economy where AI generated or impersonation tracks can exploit the name, audience and credibility of real musicians is taking shape.
“It’s very easy to do this. And now with AI… it’s not just 12 songs, it could be a thousand,” said Tewari.
Impersonation at scale
Across streaming platforms, versions of this have already begun to surface.
In late 2024, artistes who had removed their music from platforms such as Spotify found that new tracks had appeared under their names anyway. These weren’t re-uploads of the old tracks, but imitation songs, racking up streams and generating revenue.
In a widely reported case, the Australian band King Gizzard and the Lizard Wizard, which had exited Spotify on principle, found music attributed to it circulating again within months on the platform. The band had no control over these uploads and no easy way to monitor or remove them.
AI tools are enabling impersonators to attach synthetic or unauthorized content to a real artist’s identity, and in turn tapping into their audience, visibility and monetization pipeline.
“Your identity as an artist is very personal,” said Saarthak Sardana, a disc jockey and music producer. “When something shows up under your name that you didn’t create, it feels like someone walked into your house uninvited.”
Systemic issue
Unlike film or publishing, there is no central gatekeeper when it comes to music getting published on streaming platforms. The modern music pipeline is fragmented and optimized for speed.
In most cases an artiste uploads a track through a distributor, who then pushes it to multiple platforms. Once the platforms ingest the file, it attaches to an artiste profile and then surfaces to listeners. In this chain of steps, there is no mechanism to verify any of the stakeholders involved in the process.
Misattribution on music streaming platforms, either accidental or otherwise, is pretty easy, said Antariksh’s Rajput. “With DIY distribution platforms, one can easily tag another artist as a primary or secondary artist while uploading a song.”
It’s a feature built for convenience, but that’s also what makes it easy to misuse, added Rajput, pointing to a need for stronger checks.
“Misattribution has always been happening but it’s intensified with AI generated music,” said Sidhantha Jain, co-founder of M³ (M-Cube), a music marketing company that works across marketing and distribution ecosystems.
“Ever since digital distribution became the industry standard there haven’t been manual checks at scale. There is no robust identity verification,” said Jain.
“Artiste profile pages are created automatically by distributors and/or labels when they first upload a track onto Spotify for Artists,” a Spotify spokesperson told Mint in an email response. It’s not possible for an artiste to create a profile for themselves.
Until recently platforms assumed that the problem was small enough to manage reactively.
For instance, Spotify recently introduced a feature called “artist profile protection,” which allows artistes to approve releases before they go live on their profile.
“We’re currently beta testing this new feature with artistes across the globe in English who have decided to opt in,” the Spotify spokesperson said, adding that the company plans to make it available to all interested artistes in the next few months.
Rajput explains that the bigger frustration is that the cleanup is left to the artistes or their managers, who have to track down errors, and then file requests for platforms to remove fake tracks.
The default system still operates largely on trust, which is currently being strained by the scale and capability with which AI tools can now spit out musical tracks.
Tewari points out that more than 100,000 tracks are uploaded globally every day. With AI, that number is only expected to rise.
Tools such as Suno AI and Udio allow users to generate full-fledged songs, lyrics, composition, and vocals with a simple text prompt and produce music in seconds. This combination of high-volume and low-cost music production, alongside weak verification by platforms, creates an opening for impersonators.
A crook can generate dozens of tracks, upload them through a distributor, tag an existing artist as the primary performer, and rely on the platform’s systems to do the rest.
The platform’s algorithm currently sees a known artist name and pushes the track out to their listeners. Any engagement that comes through these tracks begins to accumulate on the composer’s page. And because royalties are tied to composition and publishing credits, the economic value can flow to the uploader, not the impersonated artiste. That’s because payments from streams are directed to the distributor account that submitted a track and not the profile it appears under.
Mint reached out to Suno and Udio with a set of questions but did not receive a response from the companies.
The tools behind the change
The result of AI tools such as Suno AI and Udio enabling users to generate full songs with simple prompts is visible in the numbers.
More than 106,000 tracks are now uploaded to streaming platforms every day, as of January, according to Luminate, a leading music data analytics firm.
Deezer, Europe’s homegrown streaming rival to Spotify, saw AI-generated tracks surge from 20,000 daily uploads in April 2025 to 50,000 by November 2025, representing 34-39% of all new music.
Spotify removed over 75 million “spammy” tracks, many enabled by AI, in the 12 months through September 2025, according to its latest AI protection announcement, while Apple Music flagged and demonetized roughly two billion fraudulent streams in 2025.
“We’re seeing a lot more AI-generated music in the ecosystem now,” said Rajput. “People are generating full songs at the click of a button instead of collaborating with musicians.”
For listeners and consumers, this shift is hard to detect. Rajput says even trained musicians sometimes struggle to distinguish AI-generated tracks from real ones.
Sardana also points out that most listeners care about whether a track sounds good, not how it was made.
Identity crises
Abhishek Kumar, convenor of the New Indian Consumer Initiative, a consumer advocacy platform focused on the digital economy, points out that artistes and creators in the AI age need differentiation in order to protect their art and creation.
“What we are seeing rise is a new class of actors calling themselves creators, but essentially operating as data miners,” he said, referring to those who use AI to create art.
In a series of draft representations shared with the department for promotion of industry and internal trade (DPIIT), the ministry of electronics and IT (MeitY), and the ministry of information and broadcasting (MIB), Kumar lays out how this shift is unfolding across the creative economy. He argues that the issue lies with how AI fits into the existing pipeline.
At one end are generative AI systems trained on vast amounts of creative work, often without consent or compensation. In the middle are tools that can now produce music that closely mimics a real artiste’s voice, style or identity. And at the other end are distribution platforms that allow this content to be uploaded and monetised with little friction.
Kumar describes this practice as “identity and attribution fraud” where a track can be generated anywhere, uploaded by anyone and still be attributed to a real artiste, so long as it is tagged correctly. To deal with this insurgence in AI generated music, the system, especially for identity verification, needs to be upgraded, he says.
Artistes are also calling for this shift to accommodate the changing nature of demand and supply in the music streaming industry. “At the first level, responsibility has to lie with distributors and then with platforms,” Rajput said.
In March 2026, Apple Music introduced optional “transparency tags” for labels and distributors to disclose when AI was used in artwork, lyrics, track composition or music videos.
“In cases of fraudulent uploads, manipulated streams are detected and neutralized, and royalties are not paid out for that activity,” the Spotify spokesperson clarified, declining to comment on the specifics of Tewari’s or other cases.
Mint also reached out to Apple Music, but did not receive a response from the company.
Damage beyond money
For many artists, the impact of this misattribution is deeper than just a matter of royalties.
Artistes spend years honing their craft and building a voice and loyal audience that connects with that voice. In many parts of the world and some parts of India, fan communities organize listening parties and coordinated music streaming sessions to create communities or push their favourite artistes up the charts. Now that the identity of an artist can be replicated or misused at scale, the risk becomes personal.
“We spend years building a sound, a vibe—something people recognise instantly,” said Sardana. “If that can be copied in five minutes, it messes with the whole idea of originality.”
There’s also a risk of reputational damage for musicians who spend years building their fanbase. A track that an artiste did not create can still travel to their audience and create a false perception.
“You can’t quantify the reputational damage,” Jain said. “Someone can upload a song under your name with messaging you’ve never stood for.”
This could eventually lead to a breakdown of trust between musicians and listeners.
Artistes argue that these challenges require intervention at multiple points in the pipeline, starting with stricter checks at the point of distribution and extending to faster, more reliable takedown systems on platforms.
For now, much of that burden still falls on the artiste. “The primary responsibility lies with the AI companies building these tools, because we are still figuring out what’s ethical and what’s not,” said Tewari. “Then come the platforms and distributors, everyone in the chain should be protecting the artiste’s interests.”
Tewari managed to get fake tracks in his name taken down. But the episode left a lingering question for him and others about ownership and control over their art.
In a system where music can be generated, attributed and distributed independent of them, their voices are in danger of being drowned out by artificial intelligence.
- Number of fraudulent streams flagged and demonetized by Apple Music over the whole of 2025.
- Number of ‘spammy’ tracks removed by Spotify, many enabled by AI, in the 12 months through September 2025.
- Number of tracks uploaded on streaming platforms every day, as of January, according to data analytics firm Luminate.