If you have recently stumbled upon the phrase "Adele Adelia," you are likely experiencing one of two things: either you have just watched a video that left you questioning the nature of artificial intelligence, or you have heard a song so hauntingly beautiful that you swore it was a lost demo from a major pop star.
Currently, the U.S. Copyright Office refuses to grant copyright to works generated entirely by AI. However, the producers behind argue that the composition (the piano arrangement, the mixing, the distribution) is human-made, so the entire work is protected.
Some believe that the voice is a "mash-up" generative AI model trained on two specific artists: Adele (for power and soul) and Adelia (a fictional placeholder name for a Scandinavian folk singer whose catalog was scraped without consent). The result is a vocal hybrid that no human larynx can physically produce.
Music producers have analyzed the frequency spectrum of the viral cover. They found that the vocal track contains "formants" (the resonant frequencies of the voice) that do not exist in nature. A professional singer, even one like Ariana Grande or Mariah Carey, produces formants that shift as they move their jaw. Adele Adelia’s formants are static.
If you enjoyed this deep dive into the Adele Adelia mystery, share this article with a friend who needs to know the truth behind the voice.