What Happened
An Australian producer named Josh Fawaz went from relative obscurity to the top of the national charts in a matter of months. His cover of Madonna's Like a Prayer reached No. 1 on the National Radio Airplay chart, topped the iTunes Electronic chart worldwide, and racked up 35 million streams on Spotify.
The problem, according to several music experts and working musicians, is that the song may not have been made by a human at all. They say Like a Prayer carries the fingerprints of generative AI tools, the kind that produce a finished track from nothing more than a text prompt.
Sam Whiting, a senior research fellow at RMIT's school of media and communication, told the Guardian the track is "heavily compressed" in ways that match the output of AI music generators like Suno. He pointed to what he described as "sloppy drums", vocals with "artefacts", and a low-quality streaming file as further indicators.
"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," Whiting said.
Fawaz has pushed back. "I've been doing it for 15 years and you can tell the way that AI music is produced, it has distinguishable features," he told the Guardian. He also said: "I've been releasing music way before AI was invented. What I care about [is] providing my listeners with good music. Maybe you should do the same."
Why It Matters
The dispute cuts to a question Australian broadcasters, collecting societies, and regulators have not yet answered: should artists be required to disclose when generative AI created their work?
At the moment, no such obligation exists. A track can sit at No. 1 on the national airplay chart, collect royalties, and receive thousands of radio spins without any declaration of how it was produced.
For working musicians, the royalty question is not abstract. As one musician quoted by the Guardian put it: "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."
Research published in The Conversation, drawing on analysis of more than 2 million tracks on Spotify, has raised related concerns about AI-generated or AI-assisted content crowding out human artists on streaming platforms, with flow-on effects for the royalty pools that sustain professional musicians.
Key Details
The song credits on Like a Prayer list Fawaz as the "performer", meaning the vocalist, and his uncle Fadi Fawaz (best known as George Michael's former partner) on synths and production. Neither credit addresses the use of generative AI.
APRA AMCOS, the Australian performing rights organisation, has clarified the royalty position. "The song Like a Prayer is a remix/cover of a musical work written by Madonna L Ciccone and Patrick R Leonard," the organisation stated. "As the original human rights holders of that musical work, they will be entitled to be paid all performance royalties in the usual manner."
That means the underlying composition royalties flow to the original songwriters regardless of how the cover was produced. What remains unresolved is whether Fawaz himself is entitled to performer royalties if the "performance" was generated by software rather than a human voice.
Whiting and others draw a clear line between standard production tools (Ableton Live, pitch-shifting software, AI-assisted mixing and mastering) and generative AI, where the software is the creator and the human contributes only a prompt. The distinction matters legally and ethically, but Australian law has not yet drawn that line in the context of broadcast royalties or disclosure requirements.
Background and Context
The Like a Prayer controversy is not an isolated incident. Researchers analysing Spotify's catalogue have documented a steady increase in tracks that appear to be AI-generated or AI-assisted, raising concerns about the long-term viability of royalty income for human musicians. The Conversation's analysis of more than 2 million tracks found patterns consistent with AI generation appearing with growing frequency across the platform.
Generative AI music tools have advanced rapidly. Platforms like Suno can produce a polished, radio-ready track, complete with vocals, instrumentation, and production, from a short text description. The output can be difficult to distinguish from human-made music without close technical analysis, and no reliable, publicly available detection tool has achieved consistent accuracy.
Australian radio stations have not publicly addressed whether they screen for AI-generated content before adding tracks to rotation. The song's path to No. 1 on the National Radio Airplay chart suggests that, at present, they do not.
"If this is the future of music production ... then we really are cooked," one musician told the Guardian.
What Comes Next
The immediate question is whether Australian broadcasters or collecting societies will move to require disclosure. APRA AMCOS's statement on the Like a Prayer royalties addressed the composition rights clearly, but did not indicate any policy change on AI-generated performer credits.
Regulators including ACMA, which oversees Australian broadcasting, have not announced any inquiry into AI-generated content on commercial radio. The broader question of how Australian copyright law treats AI-generated works remains unsettled, with no legislation currently before parliament to address it directly.
For Fawaz, the scrutiny has not slowed the song's momentum. With 35 million streams and a worldwide chart-topper to his name, the commercial outcome is clear. Whether the creative credit is equally clear is a question that Australian music law is not yet equipped to answer.