How AI Remixes Are Stealing Streams From Original Artists

California reggae band Stick Figure recently watched their six-year-old song "Angels Above Me" climb to number one on iTunes sales charts across six countries — but not because people were buying their version. Instead, four different unauthorized AI-generated remixes of the track went viral, racking up millions of streams that generated zero dollars for the band.

According to lead vocalist Scott Woodruff, one remix alone got 1.8 million YouTube plays in just five days. These AI remixes took the original song and changed its speed, pitch, and arrangement — keeping enough of the melody and vocals to make it recognizable, but different enough that the remix creators claimed they didn't owe Stick Figure anything.

Different Platforms, Different Rules

The band's label, Ineffable Records, has tried to remove these unauthorized tracks, but each platform handles takedown requests differently. Spotify cooperated and removed the remixes the label requested. YouTube and other platforms have been inconsistent — some comply quickly, others don't.

When the label contacted one remix creator, that person argued the remixes were "covers" (which have legal frameworks for royalty-sharing) rather than remixes of the original recording. Stick Figure's team disagreed, pointing out that true covers require proper crediting and compensation.

A Growing Problem in Music

This isn't new. Musician Steve Lacy dealt with the same issue in 2022 when his song "Bad Habit" spawned countless unauthorized sped-up remixes on TikTok. What has changed is how easy it now is to create these remixes. A few years ago, remixing required real audio production skills. Today, AI music tools let anyone generate multiple versions of a song with just a text prompt — no expertise required.

Copyright Law Playing Catch-Up

The music industry is fighting back in court. Universal Music Group sued AI music companies in December 2024, and the Recording Industry Association of America (RIAA) filed a complaint against the AI music service Udio in June 2024. The RIAA alleged that Udio can generate songs containing melodic and vocal similarities to famous copyrighted recordings when users input certain prompts.

The problem runs deeper than just one lawsuit, though. The U.S. Copyright Office began investigating AI and copyright in early 2023, but new laws haven't kept up with how fast these tools are being released to the public.

Here is the core technical problem: AI music generators learn patterns from millions of existing songs during training. They then create new audio files that may sound similar to those original songs — but because the AI generated completely new audio files (rather than copying and pasting pieces like traditional sampling), the legal question of whether this infringes copyright remains genuinely unsettled.

The Asymmetric Burden of Enforcement

This creates a lopsided situation. Creating an unauthorized AI remix takes almost no time or money. But finding and removing it across dozens of platforms takes real resources. YouTube's Content ID system — which automatically catches exact audio copies — can't detect AI remixes because they don't contain literal copies of the original recording.

Right now, the remix creators profit from ads and streaming revenue while their content sits online, sometimes for weeks before it gets taken down. Stick Figure's label has to hire people to monitor platforms and file takedown requests constantly. The economics favor the remixer, not the original artist.

This cycle has happened before in technology. When Napster let people freely share music files, when YouTube launched without clear copyright policies, and when TikTok exploded — in each case, the new platform moved faster than the law could adapt. Eventually those industries figured it out: through platform policies, court cases, and new business models. But each time, original creators took losses while the system reset.

The current AI music remix wave looks like the next chapter in that pattern. Generative AI lets people create content so quickly and easily that the old copyright tools can't keep up. How this gets resolved — whether through platform changes, court decisions, or new technology — is still being decided.

The tension runs deeper than just Stick Figure's case. As AI makes it trivially easy to create endless remix variations, attention and money will scatter across thousands of derivative versions instead of flowing to the original artists. If that continues, it could undermine the entire economic structure that makes original music creation sustainable in the first place. That outcome is not yet certain, but it is worth watching closely.