The Platform Paradox: How Spotify Became The Sheriff And Chief Beneficiary Of AI Music Fraud

by
Alisa Davidson

Printed: July 24, 2026 at 9:19 am Up to date: July 24, 2026 at 9:19 am

by Anastasiia O

Edited and fact-checked:
July 24, 2026 at 9:19 am

To enhance your local-language expertise, generally we make use of an auto-translation plugin. Please notice auto-translation is probably not correct, so learn unique article for exact info.

On July 16, 2026, Lorde opened Instagram and instructed a tech platform to maintain its synthetic intelligence away from her artwork. Spotify’s “Concerning the Tune” characteristic — an AI device nonetheless in beta that scrapes third-party sources to generate contextual blurbs — had produced a hallucinated narrative about one among her tracks, attributing a stage efficiency to the fallacious music solely. “I’m going to exit on a limb and say we don’t need this,” she wrote. “Lowering a music to an AI-generated which means proper on the supply feels prefer it limits free interpretation.”

Spotify eliminated the textual content. A spokesperson known as it a beta characteristic that might enhance with suggestions. However Lorde’s criticism was not merely a few factual error. It was in regards to the creeping colonization of music by machine-generated interpretation — the platform’s assumption that AI curation needs to be the default setting of inventive work.

For each famous person with a platform and a authorized staff, hundreds of working musicians battle a quieter, extra insidious battle. Owen Lyman-Schmidt, half of the Philadelphia gutter-folk duo Makeshift Hammer, found in 2024 that an entity known as “Carey Dupont” had stolen his band’s recordings, barely altered them, and uploaded them to Spotify underneath AI-generated album artwork and nonsense observe titles. The pretend artist had gathered 650,000 streams. The true band had 40,000.

These two tales should not separate phenomena. They’re the dual faces of the identical disaster: Spotify’s more and more fraught relationship with synthetic intelligence, and the music trade’s scramble to reply to a know-how that threatens to swallow the very notion of artistic authorship.

The Fraud Machine

This isn’t a narrative about a number of dangerous actors. It’s a story about an industrial-scale operation that has turned AI music era into what analysts estimate is a $4 billion fraud ecosystem.

Suno, a number one AI music generator, reached $300 million in annual recurring income by February 2026 and was valued at $2.45 billion. Its 2 million paid subscribers generate an estimated 7 million songs day by day — equal to Spotify’s complete historic catalog each two weeks. Most won’t ever be heard by a human ear. However a fraction are weaponized.

Fraudsters have developed their ways. As a substitute of utilizing bot farms to replay a small variety of tracks hundreds of thousands of occasions — which triggers detection — they now flood platforms with hundreds of thousands of AI-generated tracks and stream every just some thousand occasions, accumulating royalties with out elevating alarms. Apple Music confirmed it demonetized 2 billion fraudulent streams in 2025, diverting practically $17 million from reliable artists. Deezer reported that as much as 85% of streams on AI-generated music had been fraudulent, and by 2026, over 30,000 totally AI-generated tracks had been uploaded to its platform day by day.

Essentially the most brazen case is that of Michael Smith, a North Carolina musician who pleaded responsible in March 2026 within the first federal prison streaming fraud case in U.S. historical past. Smith used AI to generate lots of of hundreds of songs and deployed bot networks throughout hundreds of accounts, producing over 660,000 streams per day at peak and fraudulently accumulating greater than $10 million in royalties. He agreed to forfeit over $8 million and faces sentencing on July 29, 2026 — a date that may doubtless set the precedent for a way critically the U.S. treats AI-enabled music fraud.

However Smith was an outlier in scale, not methodology. Makeshift Hammer’s expertise reveals how the rip-off works on the grassroots degree. A bot farm scrapes recordings from platforms like SoundCloud, alters them minimally to evade detection, repackages them with AI-generated art work and near-miss titles, then cranks out automated listens. “Carey Dupont” had no web site, no social media, no live performance historical past — only a Spotify profile and 650,000 streams. When Lyman-Schmidt filed takedown notices with Spotify and SoundCloud, he obtained automated responses and, from Spotify, silence. Three and a half months after his preliminary criticism — and solely after his story was printed — did Spotify inform him the tracks can be eliminated “pending an investigation.”

The cruelty isn’t merely the theft. It’s the economics. Spotify pays artists by way of a “streamshare” mannequin: the full royalty pool is split proportionally amongst all streams. When a scammer provides 650,000 pretend streams to a pool that beforehand held 40,000 reliable ones, the identical complete payout is now break up 94% to the fraudster and 6% to the unique artist. The platform loses nothing. The fraudster income. The musician pays.

The Sheriff and the Outlaw

Spotify isn’t ignoring the issue. In September 2025, the corporate refined its AI coverage round transparency, strict prohibition of unauthorized voice cloning, and aggressive spam filtering. Within the prior 12 months, Spotify eliminated over 75 million tracks for spam and low-quality content material.

Then, on April 30, 2026, Spotify launched the “Verified by Spotify” badge — a inexperienced checkmark to assist listeners distinguish genuine human artists from AI personas. However the badge authenticates the artist id, not the content material. A verified human can nonetheless add totally AI-generated tracks. And AI-persona artists are explicitly ineligible. Spotify is drawing a line round who you might be whereas taking no place on what you make.

That is telling. Spotify doesn’t need to choose inventive benefit. It needs engagement. And AI, for all its issues, drives engagement.

This pressure crystallized on July 14, 2026 — two days earlier than Lorde’s criticism — when Spotify launched a ChatGPT-like music assistant for Premium customers. On the Q1 2026 earnings name, co-CEO Gustav Söderström acknowledged that Spotify has “the capabilities and applied sciences we’d like” to construct a derivatives enterprise on prime of current artists’ music. He didn’t disclose what these capabilities had been skilled on. Creators don’t have any disclosure framework, no audit path, and no compensation mechanism concerning whether or not their work feeds Spotify’s generative methods.

Spotify is constructing AI instruments that interpret, advocate, and doubtlessly regenerate music, whereas deploying AI instruments to detect AI-generated fraud. It’s the sheriff and the outlaw, occupying each side of a battle it helped create.

An Trade Caught Flat-Footed

The music trade’s response to this disaster has been a patchwork of legislative lobbying, authorized improvisation, and reluctant industrial lodging — and at each flip, the facility asymmetry between main gamers and dealing musicians turns into extra stark.

At GRAMMYs On The Hill 2026 in April, the Recording Academy championed three bipartisan payments: the NO FAKES Act, establishing a federal proper defending voice and likeness from AI deepfakes; the TRAIN Act, giving creators visibility into AI coaching datasets; and the CLEAR Act, mandating transparency in AI methods. The urgency behind this legislative push isn’t summary. Copyright regulation protects recordings, not voice itself — a spot that AI voice cloning has become a weapon.

In January 2026, people musician Murphy Campbell found AI-cloned variations of her voice distributed on her personal Spotify profile. The perpetrator then used the distributor’s Content material ID entry to file copyright claims towards Campbell’s unique YouTube movies, inflicting her to lose monetization on her personal content material. She described it as being “in a bizarre limbo the place I’m telling robots to take down music robots made.” Sony Music has since requested removing of greater than 135,000 songs created to impersonate artists on its roster — a determine the corporate admits is probably going a fraction of the particular quantity. In April 2026, Taylor Swift filed three trademark purposes with the USPTO — two sound marks masking her voice — to fill what legal professional Josh Gerben known as “a spot that logos could assist fill” now that AI can mimic an artist with out copying an current recording.

Whereas artists scramble for authorized safety, the foremost labels have taken a special tack. Common Music Group and Warner Music Group reached licensing agreements with AI music firm Udio, masking label-controlled rights with plans for an artist opt-in platform. The message is evident: in case you can’t beat AI, license it. However the impartial sector has pushed again. The Unbiased Music Publishers Discussion board has urged publishers to reject AI licensing agreements that fail to allocate at the very least 50% to songwriters — a warning that the financial advantages of AI offers could by no means attain the creators whose work skilled the machines.

But for all of the lobbying and deal-making, the actual energy to reshape this panorama lies not in Congress or boardrooms, however in courthouses. The complete AI music economic system is provisional, awaiting two landmark rulings scheduled for this summer season. In Sony Music Leisure v. Suno, at present earlier than U.S. District Choose F. Dennis Saylor IV in Massachusetts, a fair-use abstract judgment listening to was held in July 2026 — the primary federal case to handle whether or not coaching a generative AI mannequin on copyrighted recordings constitutes truthful use. No ruling has been issued. In GEMA v. Suno, the Munich Regional Court docket is predicted to ship a verdict on July 31, 2026, turning on Germany’s text-and-data-mining exception — a special authorized concept that might produce a divergent final result. Collectively, these selections will decide whether or not AI coaching knowledge turns into a priced licensing line merchandise or stays successfully free. If coaching is deemed truthful use, the floodgates open wider. If not, the economics of AI music era change in a single day.

However even these authorized battles, nonetheless consequential, danger obscuring the human actuality beneath the coverage debates. This can be a story about energy asymmetry, and nowhere is that clearer than in who will get heard and who will get ignored. Lorde posts an Instagram Story and will get a response from Spotify inside hours. Owen Lyman-Schmidt spends three and a half months submitting takedown notices, consulting legal professionals, and publishing an investigative article earlier than the platform acts on a transparent case of theft. Main labels negotiate licensing offers with AI corporations. Unbiased musicians get Discovery Mode, a Spotify program that lets them settle for an excellent decrease royalty fee for playlist publicity.

The platform’s structure displays this hierarchy. Spotify’s streamshare mannequin favors quantity over worth, catalog over curation, automation over accountability. AI matches this mannequin completely. It produces infinite content material at near-zero marginal value, feeds the algorithm’s starvation for novelty, and shifts the monetary danger of fraud onto the very artists the platform claims to serve. What makes this second totally different from earlier disruptions — the synthesizer, the sampler, the MP3 — is the pace and opacity of the change. AI doesn’t merely replicate or distribute music. It interprets it, impersonates it, and doubtlessly replaces it, whereas authorized and moral frameworks lag years behind.

Lorde requested for an opt-out. Makeshift Hammer requested for justice. The platform, to this point, has supplied neither. It has supplied solely extra options, extra AI, and extra silence.

Till that modifications, the ghost within the stream isn’t the AI. It’s the artist, watching their very own work disappear right into a machine they by no means requested to construct.

Alisa, a devoted journalist on the MPost, makes a speciality of crypto, AI, investments, and the expansive realm of Web3. With a eager eye for rising developments and applied sciences, she delivers complete protection to tell and interact readers within the ever-evolving panorama of digital finance.

Extra articles