GoKawiilSign up to see the future, today Can’t-miss innovations from the bleeding edge of science and tech Email address Sign Up Thank you!
The LA-based rapper Fenix Flexin is denying that his hit new single was AI-generated after extensive allegations.
The song, “Rubberz,” climbed to number 58 on the Billboard Hot 100 after dropping in June. Many of his peers gassed up the track, but listeners immediately noticed something off. A lot of things, actually.
For one, many pointed out that the vocals didn’t sound like Fenix at all, both in terms of the quality of his voice and the fact that he hits high notes he’s never demonstrated he’s been capable of in his career. His “live” performances are doing him no favor, either, as he badly lip-syncs through lyrics he barely seems to remember and which are borderline nonsensical.
Did we mention he’s singing in a British accent in the new song, for some reason, while dabbling in a genre — bad 80s synth-pop pastiche — he’s never even attempted before? Timothy “Timbaland” Mosley coming to Fenix’s defense just fueled suspicions further; he’s one of the biggest AI evangelists in the industry.
Among the multiple figures in the music industry calling out Fenix’s track was the producer Medasin, who posted a detailed breakdown on why he believed the track was AI-generated last week.
One embarrassing detail Medasin highlighted was that Fenix apparently posted a snippet of the unfinished track that had “Sonauto” in the file name. Sonauto is the name of an AI music generation software, which incidentally rebranded to Treblo just two days before his song came out.
You can also just use your years. The Verge pointed out numerous oddities in its own breakdown: The song file contains a “number of artifacts and anomalies that are commonly found in AI-generated music. Hi-hats have a certain brittle quality, the chorus vocals sound like a low-bitrate MP3, and reverb tails haphazardly cut off.”
Getting even more nerdy, songwriter and music podcaster Charlie Harding highlighted some of his own findings to the outlet.
“Ghostly backing vocals duck in and out unexpectedly. The drum reverb is full of unnatural digital artifacting reminiscent of a 64kbps MP3 downloaded off of Napster,” Harding told The Verge. “Because most generative AI companies have trained their model on low-quality MP3s of unlicensed copyrighted material, we hear audibly degraded instrumentation.”
... continue reading