AI Technology Produces Last Beatles Album Featuring John Lennon’s Voice
Although artificial intelligence (AI) has gained worldwide popularity, concerns have been raised about its negative effects by influential figures like Elon Musk and Eric Schmidt. Surprisingly, AI has also made its way into the music industry, with tools being used to replicate the voices of famous rappers like Eminem and Ice Cube, which has been criticized as “demonic” by the latter. However, this time, AI is being utilized in a unique way as The Beatles are using it to produce their final album.
Beatles member Paul McCartney told Radio 4 Today on Tuesday that they are using artificial intelligence to extract the late John Lennon’s voice from an old demo to complete the song created decades ago, adding that it will be “the last Beatles record”.
The Beatles’ AI record
According to McCartney, the idea of using artificial intelligence to create the final song came from Peter Jackson’s documentary Get Back, in which the director used custom-made artificial intelligence to extract the voices of the Beatles members from the background noise. “He could tell them apart with artificial intelligence. They could tell the machine, ‘That’s a voice. This is a guitar. Lose the guitar. And he did that,'” McCartney said.
The two surviving members of the Beatles, Paul McCartney and Ringo Starr, used source material from a demo song that Lennon had created before his death in 1980, and it was one of several that Lennon’s widow gave McCartney as a cassette. Yoko Ono in 1994.
Although producer Jeff Lynne cleaned up and released two tracks on the cassette in 1995 and 1996, a future track rumored to be “Now and Then” was scrapped due to excessive background noise on the demo track. Now artificial intelligence is being used to separate Lennon’s voice from the noise to create the song’s vocals.
McCartney said: “We were able to take John’s voice clean with this AI so we could then mix the record normally. So it gives you some leeway.”
McCartney also expressed a warning about how AI could easily copy a singer’s vocals, adding that this technology was “quite scary, but exciting because it’s the future”.
