21 million songs make up two datasets being shared by the Gen-AI community, and I’m tired of this crap

The Atlantic has been working on a series of reports about the world of generative AI and how it steals from the artists who create original material. Make no mistake. This is one element of the biggest theft, rape, and pillage of creative works in history. Every interaction, every word, every picture, and every sound is being poured into robots, and the goal for most of the corporations and governments involved is control. Control of your finances, your work prospects, and your zest for life. It competes to replace the industries that it leeches from, and cannot be considered “fair use” from that standpoint alone. AI could have been used, and often is used, as a force for good. Generative AI is inherently devoid of artistic integrity, as not a single original thought or craft went into its creation.

Alex Reisner of The Atlantic has discovered that two large datasets, one containing 12 million songs and another with 9 million, are being shared across the Gen-AI community. The datasets are enormous, and The Atlantic has compiled a database of songs in the database, which artists, composers, and performers can search through to find out if they’ve been included. Whilst The Atlantic is a paid service, they’ve kindly allowed this search facility to be available, with a more detailed example behind the paywall. My own personal music isn’t included, but my online content and videos have been used in AI training datasets elsewhere, and it makes me feel violated and angry.

Since this news was released, my Bluesky and X timelines are full of artists discovering that, in some cases, over 200 of their songs have been stolen by these AI bros. These will be the same artists looking to make ends meet because the industry has already screwed the artists over through streaming payments or predatory contracts. At some point, we have to wonder if this abusive relationship with an industry that doesn’t nurture the talent it profits from is worth salvaging. Artists deserve better. Music fans deserve better. Humanity deserves better.

Of course, this has been a slippery slope for years. Google trained its AI on 44 million tracks in 2022. Suno confirmed in 2024 that it had trained itself on, and I quote, “essentially all music files of reasonable quality” that it could download. Every time they get away with it. Currently, there are several music labels that are suing the big companies, but not a single case has ended yet. The labels, the ones that take sometimes 60%-70% of the profits an artist makes, are settling out of court to allow the likes of Suno and Udio. A nice earner that I have seen zero proof that any settlement has made its way back to the impacted artists.

It feels like the time is now to take collective action. I’m not sure who, how, or where, but we need a precedent set out in court. If their theft is allowed in court, then suddenly, the entire economic agreement on goods, transactions, ownership, and copyright goes out the window. When someone steals their model and implements it in a different way, or steals a prompt, the AI community get very upset. The hypocrisy is toe-cringingly bad. The same people arguing that “anything online is fair game” then rage when it happens to them. They fail to understand that things like copyright and paying for a working product are essential to keeping the world turning. If someone took the great AI theft to court, it would have to lose. If AI companies won, it’d be the first domino to fall for an entire economic reset. They wouldn’t be that reckless… would they?

My last thought mirrors that of Karen Hao. In her experiences with big tech, she finds that many of the leaders steering these juggernauts have a fatalistic mindset. The world is ending soon, we have maybe 30-50 years, so let’s make sure when the AI uprising or climate change catastrophe happens, we’re the rich people who can survive it in our New Zealand bunkers. They aren’t interested in long-term stability, nor AI ethics or morals. They are driven by a blindness for self-preservation, kicking the ladder out underneath them to make sure no one else can climb up. What better way to oppress people than by taking away the very things that bring them joy? If you can’t listen, watch, or feel something without questioning whether or not it is synthetically vomited out to manipulate you, then how can you trust anything in front of your eyes anymore? Isn’t that a great way to spin chaos in a world full of incompetent leaders and shady deals? If there’s only confusion, it’s very simple to create a distraction and forcefully switch narratives whilst a trade deal, or a funding spree, or a genocide, or a two-tier society ruled by tech with government in their pockets slips out the side door.

When you start thinking about it from that perspective, it starts to feel very intentional. Gen AI slop is coming for your souls, wallets, and paychecks. Stand up, unite, and resist. Fuck gen-AI.

For more information, Ed Newton-Rex has compiled a list of class action lawsuits.

Support Higher Plain Music

Higher Plain Music is part of the Higher Plain Network – a one-man indie media project. If you like what I do, please consider supporting me via Patreon for as little as $1/£1 a month. In return, you’ll receive additional perks for supporting me, such as behind-the-scenes content and free downloads. You can also donate using PayPal. Sharing the website helps too or using the affiliate buy now links on reviews. I receive a few pence per Amazon sale. All your support will enable me to produce better content, more often. I’d love to make this a full-time media network and your support can make that happen. Thank you.