Those royalties are now going to compete with AI music for a smaller slice of the pie.
In the age of algorithmic abundance, a largely unknown Australian producer named Josh Fawaz rose from obscurity to the top of national charts with a cover of Madonna's Like a Prayer — a song that music experts now suspect was shaped more by machine than by human hands. The episode is not merely about one track or one producer's credibility, but about a quiet transformation underway in the music industry: the gradual displacement of human artistry by generative systems that leave no fingerprints the casual listener is trained to notice. Australia's regulatory frameworks, still catching up to the technology, offer little recourse, while the economics of streaming ensure that every AI-generated royalty is one less coin in a human musician's pocket. What is unfolding is less a scandal than a reckoning — with what we value in music, what we demand to know about its origins, and who we believe deserves to be heard.
- A producer with a quiet decade-long career suddenly topped Australian radio and Spotify charts with an AI-suspected cover, raising immediate alarm among music professionals who could hear the machine's fingerprints in the track.
- Experts point to telltale signs — heavy compression, vocal artifacts, algorithmically smooth production choices — that suggest the credited human performer may not have performed at all, leaving song credits telling a story that may be false.
- The financial stakes are real and systemic: every streamed AI track draws from the same royalty pool as human artists, quietly diluting the income of musicians who depend on those payments to survive.
- Regulatory frameworks are a step behind — Australia's new radio code mandates transparency for AI voices on air but explicitly exempts music, leaving commercial broadcasters free to promote AI-generated tracks without disclosure or accountability.
- Tech companies are simultaneously lobbying to weaken copyright protections, seeking the right to train AI on Australian creative work, a move musicians see as a double threat: their art becomes fuel for the systems competing against them.
- The producer's evasive responses and his management's silence have left the central question unanswered, while the broader industry drifts toward a world where authenticity is optional and the origin of a song is treated as irrelevant information.
Josh Fawaz was a quiet presence in Australian music for over a decade, releasing original work to little fanfare. That changed in April when his cover of Madonna's Like a Prayer landed on streaming platforms and radio stations, climbing to number one on the National Radio Airplay chart, accumulating 35 million Spotify streams, and topping the iTunes Electronic chart worldwide. His debut album followed, charting on the ARIA Australian artist albums list with eighteen reimagined pop hits. It was a sudden, striking ascent — and it drew scrutiny.
Music experts began listening closely. Sam Whiting, a senior research fellow at RMIT's school of media and communication, identified what he described as hallmarks of AI music generators: heavy compression, vocal artifacts, production choices that felt generated rather than performed. The song credits listed Fawaz as the vocalist and his uncle on synths and production. If the vocals were machine-made, those credits described a fiction. When challenged on Instagram, Fawaz said he uses AI as a tool and that his priority is giving listeners good music. His management declined to address the production process directly. He never confirmed or denied AI-generated vocals.
The concern extends well beyond one track. A veteran DJ and producer known as Needs No Sleep went public on Instagram, calling AI-generated music the most pressing crisis in the industry today. His argument was economic: streaming royalties are a finite pool, and every AI-generated track that claims a share leaves less for human artists. He said he recognized the AI signatures in Fawaz's music the moment it arrived in his inbox through a DJ promoter service — sloppy drums, vocal artifacts, low-quality files. Fawaz dismissed the criticism as overblown.
The regulatory environment offers little shelter. A new commercial radio code that took effect on July 1st requires broadcasters to disclose AI-generated voices on air, but the rule does not apply to music. The major commercial networks that heavily promoted Fawaz's track declined to comment on their policies or explain why they aired it without investigating its origins. Meanwhile, technology companies are lobbying the Australian government to loosen copyright protections, arguing they need access to local creative output to train their AI models — a prospect musicians view as a direct threat to their livelihoods.
Sam Whiting offered a broader diagnosis: streaming culture has trained listeners to consume music passively, swiping and skipping without asking who made something or how. That inattention has created the conditions in which AI music can flourish unexamined. Commercial radio, once a curatorial gatekeeper, has chosen reach over scrutiny. The question of authenticity — once foundational to the music industry — has quietly become optional. What matters now is whether a song performs. How it was made, and by whom, has become information no one is required to provide.
Josh Fawaz was nobody in the music world until April of this year, when his cover of Madonna's Like a Prayer landed on streaming platforms and radio stations across Australia. Within months, the song had climbed to number one on the National Radio Airplay chart, accumulated 35 million streams on Spotify, and topped the iTunes Electronic chart worldwide. His debut album, Dance Like Nobody's Watching, followed with eighteen tracks of reimagined pop hits—Oasis, Cyndi Lauper, others—and charted at number eighteen on the ARIA Australian artist albums list. It was a meteoric rise for a producer who had been releasing music quietly since the 2010s, with little to show for it until he pivoted to covers with vocals.
Then the questions started. Music experts and fellow producers began examining the track closely, and what they found troubled them. Sam Whiting, a senior research fellow at RMIT's school of media and communication, identified what he called hallmarks of AI music generators like Suno: heavy compression, vocal artifacts, production choices that felt algorithmically generated rather than human-made. The song credits listed Fawaz as the performer—the vocalist—and his uncle, Fadi Fawaz, on synths and production. But if the vocals were AI-generated, those credits told a false story about who made the music.
When pressed on Instagram, Fawaz defended himself by saying he uses AI as a tool, nothing more. "What I care about is providing my listeners with good music," he wrote. He pointed out that he had been making music long before generative AI became mainstream. His management company, Hallwood, declined to answer questions about the production process. Fawaz never explicitly confirmed or denied using AI to generate the vocals, leaving the question hanging in the air.
The concern runs deeper than one song or one producer's credibility. A DJ and producer named Needs No Sleep, who has worked in the industry for fifteen years, went public on Instagram calling AI-generated music "the biggest issue in music right now." He explained that when AI-created tracks are streamed or played on radio, they generate royalties—money that flows away from human musicians who depend on those payments. The economics are brutal: as AI music proliferates, the total pool of royalties stays roughly the same, but it gets divided among more and more claimants. Human artists are competing for a shrinking slice.
Needs No Sleep told The Guardian that he could identify AI production by ear: sloppy drums, vocal artifacts, low-quality streaming files. When Fawaz's music arrived in his inbox via a DJ promoter service, he recognized those signatures immediately. He spoke up not out of jealousy but out of alarm about the direction the industry was heading. Fawaz responded by dismissing him: "It's not that deep."
The regulatory landscape offers little protection. On July 1st, a new commercial radio code of practice came into effect in Australia requiring broadcasters to be transparent about using AI-generated voices on air. But the code does not apply to music itself. Radio networks ARN, NOVA Entertainment, and Southern Cross Austereo—the major commercial broadcasters who played Fawaz's track heavily—declined to comment on their policies regarding AI music or explain why they promoted the song without disclosing its origins.
Meanwhile, big technology companies are lobbying the Australian government to weaken copyright laws, arguing they need the right to scrape Australian creative output to train their AI models. Musicians see this as a threat: their work becomes training data, their livelihoods become fuel for the systems that might replace them. The Australasian Performing Right Association confirmed that Fawaz has been a member since 2021 and that royalties from Like a Prayer will flow to Madonna and Patrick Leonard, the song's original writers, as they should. But that doesn't address the larger problem: if Fawaz is using AI to generate vocals, he is claiming credit and potentially earning income for work he did not create.
Sam Whiting framed the deeper issue this way: streaming culture and the "TikTok-ification" of music have conditioned listeners to consume songs passively, without critical engagement. We swipe, we skip, we add to playlists. We don't ask who made this, how it was made, whether it's real. That inattention has created space for AI music to flourish without scrutiny. Commercial radio, which once served as a gatekeeper and curator, has decided to promote Fawaz's track without investigating where it came from or who actually made it. The question of authenticity in music—once central to the industry—has become almost quaint. What matters now is whether the song works, whether it streams, whether it sells. The answer to how it was made, and by whom, has become optional information.
Citas Notables
This is a very impressive vocal performance if it was delivered by a human but if it's not, that brings in really worrying questions around what we value any more in terms of human expression.— Sam Whiting, senior research fellow at RMIT's school of media and communication
Those royalties that are collected are diverted away from other artists making real music. Artists count on these royalties, and they're now going to compete with AI music for a smaller and smaller slice of the pie.— Needs No Sleep, producer and DJ
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter whether Josh Fawaz used AI to make this song, if people genuinely enjoy listening to it?
Because the royalties system assumes a human made a creative choice. When you stream a song, money flows to the credited artist. If that artist didn't actually create the work, the payment structure breaks down. More importantly, it's a form of deception—listeners and radio stations are being sold a story about human artistry that may not be true.
But Fawaz says he uses AI as a tool. Isn't that just the next evolution of production, like synthesizers or drum machines were?
There's a real difference. A synthesizer is a tool you play; you make decisions about every note, every sound. Generative AI is different—you write a prompt and the system creates the output. The human input shrinks dramatically. And crucially, Fawaz hasn't been transparent about it. If he said "I used AI to generate vocals on this track," that would be one conversation. Instead, he's credited as the performer, which implies he sang it.
What about the royalty issue? Doesn't Madonna still get paid?
Yes, Madonna gets paid as the original songwriter. But the streaming pool is finite. Every dollar that goes to AI-generated music is a dollar that doesn't go to a human producer or musician who spent years developing their craft. Needs No Sleep's point is that the total pie isn't growing—it's being divided among more claimants. Human artists are being squeezed out.
Could this actually be good for music? Maybe it democratizes production, lets anyone make a hit?
That's the pitch, and it sounds appealing. But there's a cost. If AI can generate commercially viable music from a text prompt, why would record labels invest in developing human talent? Why would young musicians spend years learning their craft? The industry could consolidate around a few AI platforms, and the diversity of human expression—the thing that actually makes music matter—gets flattened.
Why haven't the radio stations just asked Fawaz directly whether he used AI?
That's the question, isn't it? They didn't ask, or if they did, they didn't care about the answer. Commercial radio is driven by metrics—streams, engagement, advertising revenue. A song that performs well gets promoted. The origin story is irrelevant to them. That's the real problem: the system has no built-in incentive to care about authenticity anymore.