Your Music Trained Suno (Now You Can Prove It)
If you’ve ever uploaded a song to the internet — any song, any platform, any year — there’s a good chance it just got named.
This week, The Atlantic published something that sent shockwaves through the music industry: four searchable databases of music used to train AI models, with 12 million tracks in one database, 9 million in another, and two more containing about 100,000 songs each. We’re not talking about vague allegations anymore. The Atlantic has put hard numbers on a question the industry has argued about for two years.
And musicians are losing it.
What The Atlantic Actually Found
The investigation, led by journalist Alex Reisner as part of The Atlantic’s “AI Watchdog” project, does something no one has managed to do before: it makes the invisible visible.
The Atlantic’s “AI Watchdog” project identified four large datasets of songs — together holding more than 20 million tracks — circulating within the AI-development community. According to The Atlantic, the collections include catalog music that “is not supposed to be free,” and they represent only a portion of the audio that developers can access to train music-generating models.
The datasets include hits from major pop artists such as Bad Bunny, Nirvana, Taylor Swift, Billie Eilish, Pearl Jam, and the Beatles, alongside jazz artists and classical composers. But here’s the thing that really matters: it’s not just superstars. Indie artists. Bedroom producers. DIY punk bands with 200 followers. They’re all in there.
Two of the four datasets have publicly documented origins, and neither was created for the purpose of training commercial music generators. The largest is LAION-DISCO-12M, a collection of more than 12 million tracks released in November 2024 by LAION, a German non-profit that compiles open datasets for AI research.
LAION, which is also behind the dataset used to train Stability AI’s Stable Diffusion image generator, says the music collection was “released for research purposes” and is intended for use “in academic settings.”
But the gap between “academic research” and commercial AI music generators apparently wasn’t much of a gap at all. All four datasets have each been downloaded several thousand times, according to The Atlantic, though because the industry keeps its training data under wraps, it isn’t publicly known which companies have used most of them.
Musicians Are Finding Themselves in the Data
The reaction from artists has been swift and visceral.
Breakcore producer Sophiaaaahjkl;8901 posted about the searchable database on social media, prompting numerous other musicians to share the post, as well as screenshots of their searches to see which of their own tracks were included in the datasets.
Backxwash, Titus Andronicus, Tre Mission, Lunice, DJ Sabrina the Teenage DJ and more are among musicians who have expressed their disdain for finding their music within the searchable database.
One musician on X posted that The Atlantic’s database showed 138 of their songs across two of the datasets — “almost my entire catalogue of music” from 2017 to 2024.
Titus Andronicus offered perhaps the most darkly funny reaction, noting that among the top six of their songs used to train Suno were “an ambient noise track” and “a deep cut from our 2022 album that no one heard or liked.”
This is the moment where AI training data stopped being an abstract legal concept and became deeply, uncomfortably personal.
The Number That Should Terrify AI Companies
Here’s why this investigation matters beyond hurt feelings: it’s evidence.
For artists, searchable proof of which songs trained a model is the hard evidence these cases have lacked.
It strengthens the labels’ cases against Suno and Udio, and it raises the pressure on platforms that still will not disclose what their own tools were trained on.
And the timing couldn’t be more legally significant. The Sony Music copyright lawsuit against Suno heads to a July 2026 summary-judgment hearing after audio fingerprinting revealed millions of copyrighted recordings in its training data.
A federal court is weeks away from a hearing that may determine whether training an AI on copyrighted recordings — and, critically, where those recordings came from — is legal under U.S. copyright law.
The key admission from Suno itself, made back in August 2024, still echoes: Suno’s training data “includes essentially all music files of reasonable quality that are accessible on the open Internet, abiding by paywalls, password protections, and the like.”
“Essentially all music files of reasonable quality” is a phrase that will end up in copyright law textbooks.
AI music companies including Suno and Udio are now grappling with at least 12 lawsuits. Sony Music is the one major label still litigating against both Suno and Udio, while Germany’s GEMA and Denmark’s Koda are also suing Suno.
Meanwhile, a separate book publishing lawsuit using piracy allegations rather than straight copyright infringement proved to be a more compelling argument, with the initial settlement pending at $1.5 billion. If the music industry follows that playbook, the numbers could be staggering.
The Licensing Split Is Already Happening
What’s fascinating is how the major labels have responded in completely different ways.
UMG and Warner have moved from litigation toward licensing, while Sony has remained in court against both companies.
Universal Music Group settled with Udio in October 2025, announcing a “compensatory legal settlement” plus new recorded-music and publishing licenses for a jointly developed AI platform set to launch in 2026.
So we’re looking at a music industry that’s simultaneously suing AI companies and partnering with them. If that sounds contradictory, welcome to 2026.
June may continue to separate the AI music market into two lanes: companies fighting over past training data and companies trying to build licensed future creation systems.
For independent musicians, this creates a genuinely confusing landscape. Your songs may have trained the very tools you’re now being asked to pay for. And no one asked your permission.
What This Means for AI Music Video Creators
If you’re a musician who uses AI tools to create visuals for your music — and that’s increasingly most of you — you’re probably wondering: does any of this affect me?
The short answer: not directly, but the ripple effects are real.
AI music generation (tools like Suno and Udio that create audio) is the target of these lawsuits. AI video generation tools — the ones you use to create music videos — operate in a separate legal space with different training data. When you use tools to create visuals for your existing music, you’re not generating new audio from potentially tainted training data. You’re adding a visual layer to music you already own.
This distinction is crucial. If you’re looking to make an AI music video from your original tracks, you’re on solid creative and legal ground. The controversy is about AI tools that generate the music itself from models trained on copyrighted audio.
That said, the broader trend toward transparency and labeling affects everyone. Article 50 of the EU AI Act requires that AI-generated images, video, and audio be marked in a machine-readable format to identify them as artificially generated. The regulation becomes fully enforceable on August 2, 2026. That’s less than six weeks away.
The risk of fines is up to €15 million or 3% of global annual turnover. For indie musicians, that fine structure won’t apply directly, but the platforms you distribute through will need to comply — and that may mean new disclosure requirements for any AI-generated content you upload.
The Streaming Flood That Fed This Crisis
The Atlantic’s databases don’t just expose what went into AI models. They help explain what’s been coming out.
Spotify said last year it pulled 75 million spammy AI tracks, and Deezer now reports that close to half of the songs uploaded to it each day are AI generated.
The Atlantic’s databases show part of what fed that flood.
This is the full circle moment: copyrighted music trained AI models, those models now generate millions of tracks, and those generated tracks flood the same streaming platforms where the original copyrighted music lives — diluting royalty pools and making it harder for human artists to get discovered.
It’s not hard to see why musicians are angry.
For context on how the streaming landscape has been shifting, check out our complete guide to AI music videos in 2026, which tracks how these changes affect visual content strategy.
Ed Newton-Rex’s FAQ for the Panicked
One of the clearest voices cutting through the noise has been Ed Newton-Rex, former VP of Audio at Stability AI who famously resigned over the company’s stance on training data. After The Atlantic’s databases went live, he posted a quick FAQ on X for musicians discovering their songs in the data.
His key points: it means it’s likely some AI music companies have used your music to train AI music models. It isn’t proof that any specific company has used it, because AI music companies generally don’t reveal what they train on. But these datasets are widely shared between AI researchers, so it’s likely your music has been used for training by at least some companies.
On the legality question: we don’t know. Lots of rights holders think it is not; some AI companies argue that it is. There are over 100 outstanding lawsuits in the US alone on the topic of AI and copyright.
And crucially, on whether it’s too late to act: no. Just because a company may have trained on your music already doesn’t mean there’s nothing you can do.
What You Should Actually Do Right Now
Let’s get practical. Whether you’re an indie artist finding your songs in these databases or a musician navigating the AI landscape in 2026, here’s what matters:
1. Search the databases
Go to The Atlantic’s searchable tool and look up your artist name. Know what’s in there. This is information you need whether you plan to act on it or not.
2. Understand the difference between AI music and AI visuals
The legal storm is around AI-generated audio. If you’re creating AI music videos for your original tracks — whether it’s hip-hop visuals, EDM content, or indie aesthetics — you’re working with AI video tools that have separate (and far less contentious) training pipelines.
3. Pay attention to tool provenance
Tool choice is becoming part of rights strategy. Where you make the song may matter almost as much as what the song sounds like. Whether you’re generating music, generating visuals, or both, understanding what trained the tools you use is no longer optional.
4. Watch the July hearing
A key summary-judgment hearing in the Massachusetts case is scheduled for July 2026 before Chief Judge F. Dennis Saylor IV. Whatever the court decides this summer could reshape the entire AI music landscape — licensing terms, tool availability, pricing, everything.
5. Focus on what you can control
You can’t un-train a model that already consumed your catalog. But you can make strategic decisions about how you use AI going forward. Creating visual content for your music is one of the highest-leverage things you can do right now. An AI music video can reach audiences on YouTube, TikTok, and Instagram in ways that audio-only releases simply can’t — and it keeps you in control of your creative output.
The Bigger Picture: Transparency Is Winning
If there’s a silver lining to this week’s revelations, it’s this: the era of AI companies training on whatever they want and hoping nobody notices is ending.
Between The Atlantic’s searchable databases, the EU AI Act’s labeling requirements hitting in August, YouTube’s AI disclosure policies, and Spotify’s ongoing efforts to identify AI content, the pressure toward transparency is coming from every direction simultaneously.
For musicians who want to use AI as a creative tool rather than have AI use them as training data, this is actually good news. The more transparent the ecosystem becomes, the easier it is to make informed choices about which tools align with your values.
If you’re looking for inspiration on how to channel AI into your visual creative work, explore our genre-specific guides — from pop music video templates to rock examples to lo-fi aesthetics. The visual side of AI is where musicians have the most creative upside with the least ethical downside.
The Road Ahead
This story isn’t over. The databases are out. The lawsuits are heating up. The July hearing approaches. And 20 million songs sit in searchable databases, waiting for their creators to find them.
What happens next will be determined in courtrooms, boardrooms, and legislative chambers. But it will also be determined by what musicians do — how they respond, how they adapt, and how they use every tool available to keep creating and reaching audiences.
The music was always yours. Now, at least, there’s proof of where it went.
Ready to create stunning visuals for your music? OneMoreShot.ai lets you generate professional AI music videos from your original tracks in minutes — no training data controversy required. Your music, your visuals, your terms.