The strange case of Fenix Flexin’s “Rubberz” has become a lightning rod in the music industry, forcing uncomfortable questions about authenticity, artistry, and the accelerating creep of generative artificial intelligence into creative fields. The track, a jarring departure from the Shoreline Mafia rapper’s signature West Coast trap sound, has climbed to number 58 on the Billboard Hot 100, but the celebration is overshadowed by a firestorm of allegations that the song is almost entirely AI-generated. Fenix has denied these accusations, but the mounting evidence—from the audio itself to the cover art to his live performances—paints a persuasive picture of a track that may have been written, produced, and performed by an algorithm.
The Sounds of Suspicion: Technical Artifacts in “Rubberz”
The most immediate giveaway is the recording’s sonic fingerprint. Close inspection reveals a suite of anomalies now widely recognized as hallmarks of AI-generated music. The hi-hats have a brittle, almost glassy quality that lacks the natural decay of a live or sampled cymbal. The chorus vocals exhibit the compressed, washed-out sound of a low-bitrate MP3, as if they were pulled from a 64kbps file downloaded decades ago. Reverb tails cut off abruptly, leaving empty space where an engineer’s careful fading would have created a smooth end. These are not the mistakes of a human producer with a modest budget; they are the predictable outputs of a model trained on degraded source material.
Charlie Harding, songwriter and co-host of the Switched on Pop podcast, emphasized the point in an analysis of the track. “Ghostly backing vocals duck in and out unexpectedly,” he said. “The drum reverb is full of unnatural digital artifacting reminiscent of a 64kbps MP3 downloaded off of Napster. Because most generative AI companies have trained their model on low-quality MP3s of unlicensed copyrighted material, we hear audibly degraded instrumentation.” This degradation is not merely cosmetic; it is a structural consequence of how these models learn. They are fed thousands of hours of compressed audio from internet archives, and the compression artifacts become embedded in the generation process.
AI Detectors Return Ambiguous Results
Running “Rubberz” through AI music detection tools does not deliver a definitive verdict. Five different detectors returned scores ranging from a 20 to 30 percent probability that the track was AI-generated or AI-assisted. But such tools are inherently unreliable. They lag well behind the generative models they are designed to catch, and their accuracy varies wildly based on training data and methodology. A 20–30 percent score is not a clear denial of AI involvement; it is simply a reflection of the detection tools’ limitations. The audio itself, as Harding and others have noted, tells a more compelling story than any black-box detector.
Beyond the Audio: Visual and Lyrical Clues
The supporting materials surrounding “Rubberz” are even less ambiguous. The single’s cover art was flagged as AI-generated with over 97 percent certainty by multiple AI image detectors. A separate post celebrating the song’s Hot 100 success received the same verdict. These numbers are not borderline; they are confident classifications. While AI image detection is not perfect, the consistency across different tools and the sheer improbability that a major label release would use obviously synthetic cover art without knowledge or intent is telling.
The lyrics, too, bear the unmistakable scent of a large language model. Producer and YouTuber Curtiss King dropped the words into Claude, where it flagged multiple signs that they were AI-generated. Rapper and critic Anthony Fantano has also weighed in, highlighting the same hallmarks. The rhymes are strikingly simplistic—clean, end-rhyme pairs that follow a rigid AA, BB structure. There are no near rhymes, no internal rhymes, no rhythmic complexity. The lines prioritize the rhyme scheme over meaning, resulting in logical near-gibberish that collapses under scrutiny.
Consider this verse:
“Swipin’ cards and stackin’ chips, I saw you sinkin’ ships / Left me standin’ in the pourin’ rain, now I bought a heavy diamond chain.”
Who is sinking ships? Are we the sinking ships? The idea is introduced and never revisited. The chorus is similarly hollow:
“Bankrolls on the table / Watchin’ local cable / Hatin’ that I’m able / Hatin’ that I’m able.”
The repetition of “hatin’ that I’m able” is awkward, and the phrase “watchin’ local cable” is bafflingly out of place in a song about wealth and success. King described nearly going into a rage spiral over that line. What is Fenix able to do? Why does he hate it? The lyrics raise questions they never attempt to answer.
The Live Performance Problem
Perhaps the most damaging evidence comes from Fenix Flexin’s own attempts to perform “Rubberz” live. Multiple in-studio appearances show the rapper awkwardly lip-syncing, at times stumbling over the flow as if he cannot remember his own song. In some clips, he appears to be mouthing along while the original recorded vocals play at full volume. The DJ strategically drops out the backing track in a futile effort to prove authenticity, but Fenix’s real voice is kept extremely low in the mix, barely audible.
When Fenix does attempt to sing live without the backing track, the results are damning. He is clearly incapable of recreating the vocal range or the fake British accent present on the recording. In one clip from a concert, he croaks out a few disconnected words while the DJ covers the gaps. In another, he stops singing entirely and lets the track play. The contrast between the polished studio version and the live reality is stark: the voice on “Rubberz” belongs to someone—or something—else.
Defending the Indefensible: Fenix’s Counterarguments
Fenix Flexin has offered a simple defense: the difference between the recording and his live vocals is down to heavy vocal processing, including AutoTune. But this argument has been widely disputed by experts. Harding, King, and Fantano have all pushed back. “AutoTune is like spellcheck,” King explained in a video. “It corrects pitch errors. It does not give you a new range. It does not give you a British accent. It does not change the way you pronounce words.” Fantano echoed the sentiment, noting that while AutoTune can be pushed to extremes for creative effect, it cannot fundamentally alter a performer’s vocal physiology. The gap between the “Rubberz” recording and Fenix’s actual singing voice is not a matter of processing; it is a matter of identity.
Fenix has also posted clips of what he claims are Pro Tools session files. But as many observers noted, those clips are as easy to fake as the song itself. Without sharing the original raw recordings and project files—a step he has not taken—the evidence remains circumstantial at best.
The Ironic Twist: “Rubberz” Succeeds Because It Sounds Real—But Is It?
One of the most ironic aspects of this controversy is the quality of “Rubberz” relative to the AI models available. Harding pointed out that if the song is indeed AI-generated, it was created with a “lesser generative AI music company with a terrible model.” More sophisticated models, he said, have already made “big improvements in the audible artifacts and awkwardly sung rhythms.” The very poorness of the song—its brittle hi-hats, its unnatural syncopation, its nonsensical lyrics—actually works against the AI explanation in some minds. “I would assume it was written by someone who has never written a song using an ’80s kids practice keyboard,” Harding said. “I am concerned that audiences find this song interesting when it is created by one of the lesser generative AI music companies with a terrible model.”
But the same roughness that might convince some listeners of the song’s authenticity is, for others, the smoking gun. No half-competent rapper, Harding argued, would ever sing the title line “I just let the rubber band snap” in such an awkward manner. “The syncopation is highly unusual, emphasizing all the wrong syllables,” he said. “By smashing so many syllables together, it forces the word ‘the’ onto the downbeat, instead of the word ‘rubber,’ forcing all of the subsequent syllables to accent with unnatural syncopation.” This is not the work of a human artist with years of rhythmic experience. It is the work of a model that does not understand prosody.
What Is the Impact on the Music Industry?
The “Rubberz” controversy arrives at a moment when the music industry is grappling with the implications of generative AI in earnest. Major labels have filed lawsuits against AI companies for training on copyrighted material. Artists from Drake to Grimes have tested the waters with AIa-simulated vocals. But “Rubberz” may be the first track to hit the Hot 100 under a cloud of suspicion so thick that even the performer’s own label seems uncertain how to handle it.
If the song is indeed AI-generated, it represents a new frontier in the blurring of authorship and authenticity. Fenix Flexin, a real human being with a real career, would be credited for a song he neither wrote nor sang. The producer credit would go to someone—or something—that does not exist as a legal person. The rights, royalties, and royalties from future performances would flow to humans who contributed little more than a brand name and a willingness to lip-sync. This is not a hypothetical; it is the reality that “Rubberz” threatens to normalize.
Moreover, the incident highlights the inadequacy of current detection methods. The audio detectors returned an ambiguous 20–30 percent probability. The image detectors were far more certain. The human ear, with trained guidance, was the most reliable instrument. But as models improve, the artifacts will fade, and the gap between detection and generation will only widen. The music industry will need to establish new standards for provenance—digital watermarks, immutable session logs, maybe even blockchain-verified recording chains—if it hopes to maintain any semblance of accountability.
For now, Fenix Flexin has done little to dispel the rumors. He has posted a few Pro Tools clips, performed a handful of unconvincing live renditions, and offered the AutoTune defense. None of it has quieted the noise. The song continues to climb, but the conversation has shifted from “Is this a good song?” to “Is this a real song?” That question may never have a satisfactory answer, but it will shape how the industry approaches AI-generated music going forward.
How Close Are We to an AI-Generated Billboard Number One?
“Rubberz” is at number 58 today. A better model, a more convincing vocal performance, a more coherent set of lyrics, and a more careful artist is likely just months away from reaching the top ten. The barriers are not technical; they are legal and cultural. The day a fully AI-generated song claims the top spot on the Hot 100 will force the Recording Academy, streaming platforms, and labels to make hard choices about definition of authorship, ownership, and artistry. “Rubberz” may be the warning shot. The question is whether anyone is ready to act on it before the shot becomes a volley.
