The Fenix Flexin AI song dispute has turned a Billboard Hot 100 hit into music’s clearest test case yet for whether a track built with generative AI can pass as human-made, and whether anyone can actually prove it either way. “Rubberz,” a synth-pop detour from the Shoreline Mafia rapper released June 5 through Flexxd Up Ent and Empire Distribution, became a genuine hit within days. It also became, almost as quickly, the subject of a public dispute that has now run for nearly two months without a confirmed resolution.
How the Fenix Flexin AI Song Dispute Started
Suspicion attached to “Rubberz” before the full song was even released. Fenix sang the track in a British accent over 1980s-inspired synth-pop production, a sharp departure from his usual Shoreline Mafia catalogue, and listeners immediately flagged the shift as a telltale sign of AI generation. Billboard reported that Fenix denied the allegations directly in the comments of an On the Radar Radio performance clip, writing that the track was “recorded same as all music I do,” and attributing the unusual sound to Auto-Tune reverb and a put-on UK accent rather than any AI tool.
The denial did not end the scrutiny. Producer Purp On The Beat, credited alongside Fenix, argued publicly that artists should be free to use AI the same way they use any other production tool, comparing it to Auto-Tune itself, without ever directly confirming or denying that “Rubberz” specifically used one.
The Evidence That Reopened the Fenix Flexin AI Song Question
Stereogum’s reporting on the dispute’s late-July escalation lays out the most detailed evidence yet. Producer and artist Medasin published a video breakdown alleging the song was built using Treblo, an AI music generator formerly named Sonauto, pointing to a since-deleted Instagram snippet Fenix had posted with “Sonauto” left visible in the file name, timed just two days before Sonauto rebranded to Treblo and “Rubberz” was released. Medasin’s video also demonstrated the software reproducing lyrical phrases closely resembling the released track when prompted with a similar style.
Fenix responded by posting what he described as project files and isolated stems intended to prove the song was built through conventional production. Medasin was not persuaded, telling his own audience that Fenix appeared to have run the AI-generated track through a stem separator and layered his own vocals on top of the machine output, an attempt, in Medasin’s account, to make the AI version look like a human recording rather than replace it with one. Fenix deleted the stems post shortly afterward. Neither side has produced evidence the other has accepted as conclusive, and Empire Distribution has not issued a public statement on the underlying dispute.
Why This Fenix Flexin AI Song Story Is Different From Past AI Music Scandals
Most prior AI music controversies involved a clear, acknowledged act: an anonymous producer cloning Drake’s voice without consent, or a label pulling a track for unauthorised use of an artist’s likeness. LiveAIWire’s earlier coverage of how generative AI is reshaping the music industry found that the industry’s legal and cultural fights so far have mostly centred on that kind of unauthorised cloning of an existing artist’s identity, a problem with a relatively clear villain and victim.
The Fenix Flexin AI song dispute is a different and, in some ways, thornier case: a credited, consenting artist potentially using AI to generate an entire track and then presenting it as conventionally performed and produced. If the allegations hold up, the harm is not to another artist’s stolen likeness. It is to listeners, streaming platforms, and chart compilers who have no reliable way to know whether what they are consuming, ranking, or paying out royalties on was made the way its credits claim.
The Detection Problem Nobody Has Solved
What makes the Fenix Flexin AI song case genuinely instructive is that, weeks into public dispute, neither side has produced evidence either a casual listener or an independent expert can call definitive. LiveAIWire’s reporting on AI-generated synthetic voice fraud found the same structural gap in a different context: detection technology lags meaningfully behind generation technology, and the burden of proving something is or is not AI-made currently falls almost entirely on whoever is making the accusation, using informal methods like file-name metadata and stem-separation software rather than any standardised, verifiable test.
That gap is precisely what has driven renewed industry interest in mandatory content provenance and labelling, the same principle behind disclosure requirements now appearing in state and platform policy. Until a reliable, independent verification standard exists for music specifically, disputes like this one are likely to be settled by public opinion and circumstantial digital evidence rather than by anything either side can prove conclusively.
The dispute also lands at an awkward moment for a music industry already grappling with a broader erosion of the line between machine-assisted and human-made creative work. LiveAIWire’s coverage of declining editorial and creative standards as AI-generated first drafts become the norm found a version of the same problem across writing and journalism: once a plausible AI output can be lightly edited and presented as fully human work, the credited author’s own account becomes the main thing anyone has to go on, and that account is exactly what is being disputed in the Fenix Flexin AI song case.
What the Rebrand Timing Actually Proves, and What It Doesn’t
The detail drawing the most attention in Medasin’s breakdown is the timing coincidence: Sonauto AI rebranded to Treblo two days before “Rubberz” was released, and a snippet Fenix had posted earlier still carried “Sonauto” in its file name. That timing is suggestive, but it falls well short of proof on its own. A shared file name naming convention does not establish that the released track itself was generated by the software, only that the software’s name appeared somewhere in Fenix’s files around the same period. Fenix’s own stems, whatever their ultimate origin, have not been independently examined by anyone outside the dispute itself, and no music forensics lab or platform has issued a public finding either way.
That evidentiary gap is the story as much as the accusation is. A dispute this prominent, involving a charting single from an artist with an established fanbase, would be an obvious candidate for independent verification if the tools and institutional will to provide it existed. Their absence says as much about where AI music detection currently stands as anything Medasin or Fenix have posted.
What This Means for You
If you are a listener trying to work out whether a song you like was AI-generated, the honest answer is that no consumer-accessible tool currently gives a reliable verdict, and the file-name and stem-separation evidence driving disputes like this one requires technical expertise most listeners do not have and platforms do not currently verify before a song charts. Streaming platforms and performance rights organisations have not yet built the kind of provenance-checking infrastructure that would let a listener, or a chart compiler, confirm authorship with confidence before a track earns airplay, chart position, or royalty payouts.
The more durable lesson sits with the credits rather than the melody. As AI music tools become cheaper and more capable, the credited artist’s public statement about how a track was made is likely to remain, for now, the only account most listeners ever get, whether or not that account turns out to hold up under the kind of scrutiny “Rubberz” has now received for the better part of two months.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity, and the social impact of emerging technology. He publishes daily at LiveAIWire.com.