When it comes to creating music with AI and machine learning, there is quite a bit of controversy. There are pressing questions about whether AI-produced music should be considered original or not. If so, it’s not clear who may have the rights to that music, especially when AI mimics another artist’s style. But despite these issues, there appears to be some uses of AI in musical production with which few have a problem. Recently, AI was used to helped with The Beatles last song released titled “Now and Then.” In the song, the actual voice of John Lennon is used thanks to these new technologies to create the new song. And once you hear it, it becomes clear just how impactful AI will likely be in our lives.
The Beatles last song released is a project that actually began nearly three decades ago. After receiving a recording of John Lennon working on a new song in 1980, the remaining Beatles saw an opportunity. Unfortunately, their efforts at that time faced obstacles because the technology wasn’t yet available to isolate sounds. But that has since changed, which prompted Paul McCartney and Ringo Starr to pursue the project once again last year. Leveraging machine learning abilities, they were able to combine inputs from all four Beatles’ members into the song. In this regard, The Beatles last song released is truly a collaboration. Creating music with AI in this capacity is far less controversial than other pursuits. And it also demonstrates that AI is likely to be a major disruptor to the industry whether we like or not.
The Beatles Last Journey Together
The Beatles last song released will be part of a double A-side single that’s paired with the band’s first song ever. “Now and Then” will be on one side while their 1962 debut song “Love Me Do” will be on the other. But the journey to get to this point was a long one. A conversation between Paul and Yoko Ono took place in 1994 that prompted it. Yoko Ono revealed that she had a recording of Lennon working on a song in 1980 right before his murder. After receiving the tape from Yoko Ono, Paul approached Ringo and George Harrison about creating a new song using John’s voice. But despite valid attempts to do so, they were unable to dissect John’s voice from the background piano. As a result, the idea was tabled.
That might have been the end of things had it not been for the making of the band’s recent documentary. Producer Peter Jackson released The Beatles – Get Back in 2022. And instrumental to its making, Jackson used machine learning and AI to refine the sounds and music. These technological advances actually used AI to carve out different sounds from a track. Thus, it became possible to better isolate Lennon’s voice from his 1980’s recording. Now that this could be accomplished, it was feasible to produce a new Beatle’s song using John’s actual voice. And they were able to also use Harrison’s work on the song from 1994 since he passed in 2001. Though not the same as the band’s collaboration in person decades ago, the Beatles last song released still has input from all of them.
Change Resistance in the Industry
The Beatles last song released is a beautiful melody that shows the benefits of creating music with AI. Without such technology, Lennon’s voice could not have been captured well enough for the song. But that doesn’t mean many musicians aren’t displeased with the use of AI in the industry. Singer Nick Cave has labelled the use of AI to mimic his style as a grotesque mockery. Bad Bunny has been a highly vocal critic of creating music with AI replicating his voice with his fans. Their concern is that AI-produced music that imitates existing singers’ music or voice infringes on their creative rights. But this is a highly gray area right now that has yet to be sorted out. The only difference between the Beatles last song released and AI used in these instances involves artist approval.
Many platforms now exist for creating music with AI that are either open source or readily available to the public. An additional concern that many artists have related to these platforms involves how AI is trained. By combing the Internet, AI platforms creating music are trained from published songs, lyrics, and vocal styles. But because the resulting music produced via AI is unique and original, true copyright infringement is said not to exist. The question is who actually has rights to AI creations. Some argue that those creating music with AI should have copyright ownership since they are simply using AI as a tool. But others believe no ownership exists at all since AI did all the creating. Given these controversies, it’s not surprising many musical artists oppose the infusion of AI into the industry.
A Threat or an Opportunity?
While many musicians are threatened by AI, there are others who are embracing its potential. For example, Taryn Southern has been creating music with AI before adding her own voice and melodies to the product. Holly Herndon created a deepfake of her own voice using AI in 2021 called “Holly+.” She allows other artists to use her voice in their own musical creations through this platform. And Grimes launched Elf.Tech, which allows others to use her voice in their music I exchange for song royalty fees. In this regard, it’s clear that some artists see opportunities in creating music with AI. Without a doubt, the Beatles last song released adopts this perspective as well.
Without question, there are some issues that need to be resolved when creating music with AI. But its use in bringing Lennon’s voice back to life in the Beatles last song released shows its positive potential. Whether it’s related to content creation, music, or art, machine learning and AI have arrived. Resisting these trends will therefore be quite futile for creators moving forward. As such, it’s perhaps best to follow in the Beatles’ footsteps and find innovative ways to leverage AI for good. Using AI as a new creative tool could very well be the chance many new artists are seeking.