AI Slop Is Flooding Music Streaming Apps

AI-generated songs are flooding streaming platforms and forcing new rules on labels, verification, and royalties.

In 2025, Luminate says 106,000 songs were uploaded every day across streaming and other music platforms. That volume helps explain why tracks built from AI clones and near-copy remixes can spread before anyone notices.

That is the core warning in The Atlantic piece by Will Gottsegen: the music business is entering a phase where AI-generated tracks can look, sound, and chart like human work, while the platforms that host them are still trying to catch up.

How AI remixes started slipping through

The article centers on a wave of songs with names like “Angel Above Me” and “Run Run River,” which spread on Spotify and TikTok and even reached No. 1 on iTunes in Germany and Austria. Those tracks appear to be AI-generated versions of “Angels Above Me,” a 2019 song by Stick Figure.

That matters because the new versions are not simply covers. They blur the line between derivative work and copied work, and in some cases the original co-writers are not credited. For listeners, the result is a playlist problem. For artists, it is a credit and payout problem.

Gottsegen points out that AI music is not the issue by itself. Some artists already use AI tools in their own workflow, and that can happen within legal and commercial rules. The real mess starts when a system can generate a track that sounds close enough to a known song to ride the algorithm without permission.

• Some tracks gained millions of streams on Spotify and TikTok.
• Versions of the same song reached chart positions in multiple countries.
• DIY distributors can push songs to major services for a small fee.
• Not every upload gets reviewed closely enough to catch infringement.

Why the legal tools are not enough

Music law already has a few well-worn paths for borrowing. Parody is allowed in many cases, covers are legal with the right license, sampling can work with clearance, and interpolation also has rules. That framework is built for identifiable human choices, not for automated systems that can mass-produce near matches in seconds.

The article gives a simple example: Ariana Grande’s “7 Rings” reused a tune from The Sound of Music, and the rights holders took 90 percent of the songwriting royalties. That is the normal world of licensed reuse. AI remixes sit outside that world when they are made without permission and uploaded at scale.

Gottsegen also describes how fast the tools are. He says he could generate an AI clone of Kendrick Lamar’s voice for a diss track in about 30 seconds, and could just as easily create an Elliott Smith-style song with the original lyrics. He did not upload them, but the point is obvious: the barrier to making plausible copies is now very low.

“It’s not some coming danger; it’s already happening,” Will Gottsegen writes in The Atlantic.

That line lands because the industry’s current defenses are mostly reactive. The Digital Millennium Copyright Act can remove infringing tracks one by one, but it does not stop the flood. If 10,000 copies appear, a takedown process that works one song at a time becomes a cleanup crew, not a wall.

Platforms are moving toward proof of human work

The more interesting shift is not just about spotting fake music. It is about proving what is real. Instagram chief Adam Mosseri wrote in December that as AI gets better at imitating reality, “it will be more practical to fingerprint real media than fake media.” That is a big admission from a platform that already hosts a lot of synthetic content.

OpenAI CEO Sam Altman is backing a startup that scans eyeballs to produce proof of humanness, which shows how far identity verification is spreading beyond music. On the music side, Spotify has begun rolling out an artist verification badge for accounts that meet its authenticity criteria.

The catch is that these systems can exclude purely AI-generated artists while still allowing people who use AI tools “responsibly.” That leaves a huge gray area, and the rules are still being written while the uploads keep coming.

• Spotify says it removed over 75 million spammy tracks in the past year.
• Spotify also says it withholds royalties from manipulated streams.
• Spotify’s new badge system is tied to authenticity criteria.
• Instagram’s view is shifting toward verifying real media, not labeling every fake.

Why TikTok, Spotify, and labels are all tangled together

The article’s sharpest contradiction is business-driven. Last Friday, Universal Music Group announced a licensing deal with TikTok that expands protections against AI music. A day earlier, Universal had announced a partnership with Spotify that allows users to create AI-assisted remixes of selected songs.

That is the tension in one sentence: the same companies want AI controls and AI features at the same time. Spotify co-CEO Alex Norström said the remix deal is “grounded in consent, credit, and compensation for the artists and songwriters that take part.” That sounds good, but the real test is whether the permissions are actually enforced when uploads start multiplying.

There is also a deeper cultural issue. Spotify’s business model has long depended on passive listening, especially in the form of endless “chill” playlists. As Liz Pelly argues in Mood Machine, the service profits when music fades into background noise. That habit makes it easier for synthetic tracks to blend in unnoticed.

The article ends on a hard truth: if listeners are not paying close attention, AI-made copies can earn streams, siphon royalties, and distort credit before anyone realizes what happened. The next question for Spotify, TikTok, and the labels is not whether AI music exists. It is whether they can build a system that rewards consent fast enough to outrun the spam.

My bet: the winning platform will be the one that can verify human authorship at upload time, not the one that promises to detect every fake after it goes viral.