top of page
Search

Can We Ever Trust Music Again? Inside the Collision Between AI and Authenticity

  • Writer: James Garner
    James Garner
  • Dec 18, 2025
  • 9 min read

Updated: Dec 19, 2025

Imagine walking into a bar and hearing your favourite song. The production is flawless, the arrangement ideally suited to the venue's atmosphere. You close your eyes and let the melody wash over you. Then someone leans across and says, "That's AI." Your eyes snap open. Does it feel different now? More hollow? The same, but somehow less real?


This is the dilemma facing the music industry in 2025, and it's far more nuanced and troubling than most of us realise. Two accomplished musicians, drummer Anna Mylee and songwriter Tim Fraser, have been grappling with this exact question. Their insights reveal a problem that extends far beyond concerns about technological displacement. It touches on something more fundamental: what happens to authenticity, artistry, and the very soul of music when machines can create work indistinguishable from human creation?


What emerges from their analysis is not a simple story of technology threatening livelihoods. Instead, they paint a complex portrait of an industry already hollowed out by decades of poor decision-making, now facing an accelerant to its own decline.





The Moment Everything Changed

Anna Mylee remembers the exact moment she realised AI would fundamentally transform music. It wasn't when she first heard an AI-generated song. It was when she first learnt that AI-generated music even existed.


"I was in front of my TV, and I said, That's it. That's it. That's over," she recalls. "Just waiting for it to happen."


What struck her most wasn't the technology itself, but its implications for trust. If machines could generate music indistinguishable from human work, how could we ever honestly know what we were listening to? This wasn't just an economic threat to musicians. It was an existential one to music as an art form.


This realisation sent Anna Mylee down a research rabbit hole. She created a video series exploring how AI was reshaping the music industry, and the deeper she dug, the more troubling the picture became. Her initial optimism that AI might push audiences back towards genuine, live performances and original music gave way to something darker: the recognition that even "authentic" work would be treated with suspicion. Even if someone claims AI didn't create their music, how do we ever verify that?


This concern has real practical implications. YouTube videos uploaded years ago are now being remastered through AI without the original creators' knowledge or consent. There is no notification to the uploader. There is no permission requested. Their content becomes hybrid, and they remain entirely unaware.


A Paradox at the Heart of the Industry

Here's where Tim Fraser offered a perspective that reframes the entire debate. He didn't dismiss the threat of AI. Instead, he pointed out that the music industry's vulnerability to AI is simply the latest chapter in a much longer story of abandonment.


Tim traced the industry's capitulation back to the adoption of Spotify in the UK. When music executives accepted Spotify's payment model, they made a choice. They chose short-term survival over long-term principles. Industry leaders described the deal as "better than nothing." This reasoning became embedded in the industry's DNA.


So when Universal Music Group signed a partnership with OpenAI and Disney committed $3 billion to make its material available for AI training sets, the logic was identical. As long as the money keeps flowing to established power brokers, the mechanism by which that money is generated becomes irrelevant.


Anna put it bluntly: "The only person not making money is the artist. But the artist, we haven't cared about whether the artist is getting money or not for a long time. That's not new."


The three major record labels have essentially endorsed AI by signing partnerships with technology companies. A few independent labels have taken a stand. Pelagic Records, which specialises in metal music across Belgium and Germany, declared that it would never engage with AI. Some radio stations have adopted similar policies. But when the industry's most prominent players say yes, their decision tends to set the market direction.


The Copyright Collapse Nobody Talks About

One of the most unsettling aspects of the conversation involved the erosion of intellectual property protections, long before AI arrived.


The discussion touched on interpolation: when artists take a recognisable melody or portion of a song and incorporate it into their own work without explicit credit or permission, unlike sampling, where producers typically clear rights and credit original artists, interpolation exists in a murky legal grey area. And it's becoming increasingly common in mainstream pop music.


Anna Mylee described a particularly telling example: when Nicki Minaj interpolated Tracy Chapman's classic without permission. Chapman sued and initially lost but won on appeal, a rare moment when an artist successfully defended her creative territory. "She made a point," Anna explained. "It's like this is my song. I have the right to know and to want or not want something done with it."


The principle is clear: if you want to cover someone's song, that's an entirely different matter. You can do that freely. But if you take a substantial portion of a melody, incorporate it without credit, and don't secure permission, you've blurred the line between covering and stealing. Tracy Chapman's victory established this boundary. Yet interpolation continues to grow in pop music.


Here's what makes this relevant to AI: the industry has been gradually normalising the idea that other people's creative work can be repurposed, reused, and recontextualised without meaningful consent or compensation. AI didn't invent this problem. It simply accelerated and automated it. Everything AI-generated music has been trained on is, under current copyright law, considered stolen work. But the industry has already been stealing in slow motion for decades.


The Authenticity Question That Can't Be Answered

The heart of the conversation kept returning to one impossible question: what makes music authentic?


One perspective emphasises the importance of personal story. When we connect with an artist, it's often because we know what they've been through, their struggles and triumphs. This knowledge gives the music emotional resonance. But is this the only path to authenticity?


Tim introduced a complication that destabilised this certainty. He pointed to Randy Newman, one of his heroes, who has spent his entire career writing about characters and situations entirely removed from his own life. Newman doesn't pour autobiography into his work. His songs explore racism in the South, create characters for Toy Story, and conjure worlds that have nothing to do with his personal journey. Yet his songs are no less profound, no less worthy of our emotional investment.


This opened a crack in the argument. Perhaps authenticity isn't about the artist's personal journey. Maybe it's something more subtle: the presence of conscious creative choice, deliberate artistic intent, and the fingerprints of a particular human sensibility.


Which raises an uncomfortable question: if an AI system produces music that a talented songwriter then refines, manipulates, and transforms into something new, is that authentic?


Tim seemed to suggest it might be. He uses AI to create sonic versions of songs he's written, transforming £500 demos into polished £8,000-quality versions without changing a single chord, melody, or lyric. The songwriting remains his. The AI provides the gloss. This allows him to focus his resources on the compositions he genuinely believes deserve investment and development.


But Anna's counter-argument cuts deeper. The real problem emerges when someone with no musical knowledge, no years of study and sacrifice, can produce work of comparable quality to a lifelong musician in a matter of seconds. "Why?" she asked. "Because in a prompt, in a few seconds and with no knowledge, you create something of the same quality, not the same value, but the same quality."


The distinction she's making matters. Quality and value are not identical. A technically competent AI composition and a human-created song might sound equally polished. But one contains genuine artistic intent, conscious creative choice, and the fruits of dedicated study. The other contains none of these things. Yet to the ear, they may be indistinguishable.


Two Futures Emerging

What became clear as the conversation progressed is that we're not facing a single future for music. We're facing two diverging paths.


On one path, music is pure consumption. Entertainment without artistry. Functional background music for cafes and bars. Generated music is so cheap to produce that paying humans to compose becomes economically irrational. This is already happening. Anna mentioned that she's noticed AI music appearing in restaurants and shops precisely because venues don't have to pay PRS (Performing Right Society) royalties.


The BBC has also begun broadcasting AI-generated artists on BBC Introducing, representing an institutional endorsement of machine-created music.


On the other path sits music as art. Human-created work is valued specifically because it emerges from human experience, human struggle, and human creativity. This world already exists. Vinyl has experienced a genuine resurgence. Artists are finding audiences who value authenticity enough to support them directly. Younger listeners are actively seeking out music from before the AI era, valuing it precisely for its human origin.


Tim speculated that record labels will eventually split along these lines. Major labels will embrace AI-generated performers, constructed through market research to appeal to demographic cohorts. Independent labels, specialising in specific genres and committed to human artistry, will command premium prices and deeper fan loyalty.


The irony is that this split might actually be healthier than the current streaming monoculture. At least it would be honest about what music is: in one world, it’s entertainment. In the other, it’s art.


When Does AI Deserve to Exist?

An interesting tension emerged when considering whether AI has legitimate applications in music at all. Tim introduced a crucial distinction that applies across industries: some uses of AI deserve to happen precisely because they address market failures.


If AI can reduce the cost of writing a will or a house conveyancing document, opening these services to people who couldn't previously afford legal help, then perhaps that's a net social good. These are "closed shops" that deserve to be opened. Solicitors have long overcharged for routine legal work simply because they held a monopoly on the service.


Similarly, there are music contexts where AI might serve a genuine purpose. Jingles for podcasts. Background music for video games. Music for television quiz shows, which Tim noted often rely on simple sound effects and cues rather than complex compositional work. In these spaces, paying human composers significant fees for functional, repetitive work may not be justified.


Yet music that stands alone, music created for its own artistic value, music intended to move and transform the listener, occupies different territory. It loses something essential when the human is removed from the equation.


The Genie's Out of the Bottle

Neither Anna nor Tim was naïve enough to suggest AI can be stopped or meaningfully regulated. Tim's metaphor was stark and historically grounded: "This is like the Manhattan Project. We've had the first nuclear explosion, and it's here, and we can't turn the clock back now."


The question isn't whether AI will continue to advance. The question is how we respond to it. But both found small reasons for hope. Anna envisions a future where artists return to creating music primarily for themselves, for the love of it, rather than as a commercial proposition. Perhaps a secondary income funds the artistry, allowing musicians to create without compromise. This represents a return to an older model of music-making, where artists supported themselves through teaching, performance, or other work while creating their art.


Tim's hope rested partly with music consumers who instinctively value authenticity and are willing to seek it out. He cited Terry Britton, a songwriter who recorded his guitar and vocals using AI production tools, creating something "sublime." The principle was simple: rubbish in, rubbish out. Great song in, great song out, but glossier. When the foundational creativity remains human, AI can serve as a magnifier and enhancer.


Perhaps the future isn't AI replacing musicians but serving as a tool in the hands of musicians who understand what they're doing and maintain conscious creative control.


Why This Conversation Matters Beyond Music

What makes this discussion resonate far beyond the recording industry is its universality. Every profession has identical conversations about AI. A radiographer with 30 years of experience wonders why their expertise matters when an algorithm can produce results in seconds. A surveyor contemplates a future where AI interprets data faster than any human could. A solicitor considers how AI might handle routine contract work.


The fear is the same across professions. The displacement is identical. The question is unanswerable: if technology can do something faster and cheaper, how do we justify continuing to pay humans to do it?


The music industry faced this question earlier and more visibly than most. But the underlying anxiety is spreading across the professional landscape.


Listen to the Full Conversation

The insights above capture the essential tensions and themes, but they only hint at the richness of what transpired in this conversation. The whole episode explores the Beatles' use of AI technology to recover John Lennon's voice from archival material and asks whether this represents an ethical use of AI for creative restoration. There's an extended discussion of whether selling something like Kiss's identity rights to enable future AI-generated albums is a natural business decision or a troubling capitulation.


Tim offers a withering critique of BBC Introducing's decision to broadcast AI-generated music, questioning why a publicly funded institution would endorse this path. Anna discusses her second video project, which took on a darker perspective than her first, reflecting the deepening concerns she encountered through her research.


Perhaps most valuable are the unresolved moments: the genuine uncertainty about what authenticity even means in 2025, the wrestling with how to maintain one's craft as the industry's foundations shift, and the recognition that this challenge extends far beyond music into every professional domain.


Listen to the full episode on Project Flux to hear how these accomplished musicians are thinking about the future of their craft and what we might lose in pursuit of convenience and efficiency.


Links and Stuff:


Tim Fraser






Lulu is doing another fab Tim song Could I Be More Blue https://open.spotify.com/track/0Dj7Nuk4jMhAoUWIFwOUB9?si=9d8187475c7d49ed


Anna Mylee












 
 
 

Comments

bottom of page