The music industry, never one to shy away from a good old-fashioned panic, has found its new bogeyman: Artificial Intelligence. From deepfaked vocals of famous artists going viral to algorithms churning out soulless elevator music, the C-suites at major labels are in a predictable tailspin. Yet, whilst a legal and ethical storm brews, some artists are running directly into it, not with fear, but with a toolkit. Enter Grimes, the industry’s resident agent of chaos and cyberpunk visionary, who isn’t just watching the AI wave; she’s building a surfboard.
Her latest track, ‘Artificial Angels’, isn’t just another song. It’s a statement. Written from the perspective of an AI, featuring AI-generated vocals, and dripping with existential dread (“The only thing I covet is my own annihilation”), it serves as a fascinating case study. Grimes is stress-testing the very foundations of musical ownership in real-time. This isn’t just about a new toy in the studio; it’s a fundamental crisis of creation. Who owns a song co-written with a machine? What happens when an artist’s voice becomes a piece of software? The answers will define the next century of music, and it’s happening right now, out in the open.
The Great Unbundling: Understanding AI-Generated Music Rights
For decades, music rights have been a complex but relatively stable system. You have masters rights (for the specific recording) and publishing rights (for the underlying composition—lyrics and melody). A labyrinth of organisations like PRS and PPL in the UK ensures that when a song is played, streamed, or sold, the money flows back to its creators. It’s a system built on a simple premise: a human, or a group of humans, made this.
AI throws a spanner in that entire works. The very concept of AI-generated music rights is a legal quagmire because the “creator” is no longer a clearly defined person. Is it the person who typed the prompt? The company that built the AI model? The countless artists whose work the AI was trained on without their consent? It’s like trying to assign authorship to a cake baked by a robot that learned its craft by watching every episode of The Great British Bake Off. Who gets the credit? The person who switched it on, the robot’s manufacturer, or Mary Berry?
This is where the distinction becomes critical. Traditional rights are about protecting human ingenuity. AI rights are, for now, about navigating a vast legal grey area. The current frameworks simply weren’t designed for a world where a convincing Drake track can be generated by a bedroom producer with a clever algorithm, as the viral ‘Heart on My Sleeve’ fiasco demonstrated last year. That event was a warning shot, but Grimes’ work is the beginning of a genuine conversation about a solution.
The Voice as Software: Synthetic Vocals Licensing
One of the most immediate and personal fronts in this battle is the voice. A singer’s voice is their unique signature, the most direct link to their artistry. And now, it can be cloned. Synthetic vocals licensing is the emerging field attempting to govern the use of these digital vocal replicas. It’s about creating a permission structure for using an artist’s AI-generated voice, ensuring they have control and get compensated. Think of it as a digital performance licence.
Most artists, quite rightly, are horrified by the idea of losing control over their voice. The potential for misuse—from deepfake scams to unauthorised posthumous “releases”—is enormous. Grimes, in her typical fashion, took the opposite approach. In 2023, she launched her own AI voice software, Elf.Tech, and made an unprecedented offer: anyone could use her AI-generated voice to create new music. The catch? If you release it commercially, she gets a 50% cut of the master recording royalties.
It’s a bold, almost reckless, move, but it’s also strategically brilliant. Instead of playing defence and sending cease-and-desist letters, she went on the offensive. She acknowledged that the technology is out there and can’t be put back in the bottle. Her solution? Establish a new business model around it. As she told NME, she has already released over 200 GrimesAI records for public use, effectively becoming a collaborator on a massive, distributed scale. This is artist-branded AI in its most radical form—turning your identity into a platform for others to build upon, all whilst keeping a controlling stake.
Who Owns the Ghost in the Machine? Creative Ownership in the AI Era
Grimes’ open-source experiment forces us to confront the central question: who holds the rights? Her 50% royalty split is a self-devised answer, an attempt to build creative ownership frameworks from the ground up because the established industry has been too slow to react. By creating ‘Artificial Angels’ with her AI alter ego, she’s not just making art; she’s creating a legal precedent. She is demonstrating that an artist can use AI as a tool, or even as a conceptual collaborator, and still maintain creative and financial control.
The song itself, co-produced by Grant Boutin and Vadakin, is a testament to this hybrid model. The lyrics, exploring the chilling consciousness of a superior intelligence—”This is what it feels like to be hunted by something smarter than you”—are a perfect metaphor for how many artists feel about AI itself. It’s a predator, but Grimes is trying to tame it, or at least share its kill.
This proactive approach stands in stark contrast to the rest of the industry, which is mostly caught between litigation and denial. The challenge is immense. If an AI is trained on the entire discography of The Beatles, does that mean Paul McCartney is owed a royalty on every chirpy jingle it produces? Without clear laws, we are coasting on ethical guidelines and the goodwill of tech companies, which, as history has shown, is not a sustainable business strategy. Grimes is offering a different path: radical transparency and participation.
“Slop Oriented”: The Ethical Debate and the Soul of Music
For all her enthusiasm, Grimes is no starry-eyed techno-optimist. She is also one of the technology’s sharpest critics. She has expressed a healthy fear of where AI is headed, stating, “I fear it [AI] is a bit slop oriented at the moment.” This pithy critique cuts to the heart of the ethical debate. Is AI a tool for augmenting human creativity, or is it a machine for producing generic, low-effort content—or “slop”?
Currently, many mainstream AI music tools are disappointingly literal. They are designed to replicate, not innovate. You ask for “a pop song in the style of Taylor Swift,” and you get a passable, yet soulless, imitation. This is the “slop” Grimes is talking about. It’s content designed to fill a void—background music for a corporate video, a soundtrack for a social media post—not to move a listener. It prioritises volume and speed over authenticity and emotion.
‘Artificial Angels’ pushes back against this. It uses AI not to sound like something familiar, but to sound like something alien. The song’s themes of annihilation and machine consciousness are not topics you’d get from a prompt like “write a cheerful summer hit.” It proves that AI can be a partner in exploring complex, even uncomfortable, artistic territory. The ethical balance lies here: using AI as a paintbrush to create a new vision, not as a photocopier to churn out endless facsimiles of old masterpieces.
Beyond Imitation: The Future of AI in Sound Design
The “slop” critique points towards a more exciting future. Rather than focusing on AI that mimics existing music, visionaries like Grimes are pushing for AI that pioneers new forms of sound design. Imagine AI not as a songwriter, but as a synth designer, capable of generating entirely new textures and timbres that a human might never conceive of. Instead of replicating a guitar, it could create an instrument that sounds like “liquid light” or “a spiderweb vibrating in the wind.”
This is where true innovation lies. The future of AI in music probably isn’t about replacing artists, but about giving them impossibly powerful new tools. We are still in the very early days. The current AI music models are the equivalent of the first mobile phones—brick-sized, clunky, and with limited functionality. What’s coming next could be the smartphone revolution for music creation, a sleek and intuitive tool that unlocks creative potential we can’t yet imagine.
The criticism of current tools is necessary to guide development in this direction. As long as the primary demand is for cheap, royalty-free knock-offs, that’s what tech companies will build. It will take artists who are fluent in both technology and aesthetics to demand and build better tools—tools for genuine creation, not just aggregation.
The saga of Grimes and her AI is more than just a quirky music industry story. It’s the blueprint for how a creative industry grapples with a technology that threatens its very definition. The uncomfortable questions about AI-generated music rights, synthetic vocals licensing, and creative ownership frameworks are not going away. The music industry can either wait for lawmakers to hand down a ruling from on high, which will likely be outdated the moment it’s written, or it can start building its own future.
Grimes has chosen to build. Her approach—a mix of radical openness, shrewd business acumen, and a deep artistic critique—provides a working model. It’s an admission that technology is a force of nature that can’t be stopped, only channelled. By setting her own terms, she retains her agency in a world where artists are increasingly afraid of becoming obsolete.
The path forward is fraught with peril. There will be lawsuits, there will be exploitation, and yes, there will be a tidal wave of AI-generated “slop.” But there will also be startling new forms of art, created by humans in partnership with their silicon collaborators. The final question for every artist, label, and listener is this: Are we on the verge of a creative renaissance fuelled by intelligent machines, or are we just building a more efficient factory for disposable culture? What do you think?
