Some 60 years after Beatlemania, The Beatles were back in the former USSR—and just about every other country around the world—in a big way last week, thanks to their first “new” song since the mid-1990s.
“Now and Then” uses AI-led technology called MAL (machine audio learning) to flesh out a John Lennon piano demo, ultimately turning it into a haunting ballad that has amassed 8.3 million plays on Spotify since its Nov. 2 debut. There have also been more than 20 million views of the official video on YouTube.
With Lennon’s unmistakable voice front and centre, accompanied by sweeping strings, piano, a George Harrison-esque guitar solo (played by Paul McCartney, who faithfully replicates his old bandmate’s signature soaring sound) and typically understated drumming from Ringo Starr, “Now and Then” sounds uncannily like something from the Beatles’ Let it Be era, or Lennon circa “Imagine.”
As a huge Beatles fan, it’s great to get any version of the band back into my life. But the song’s creation has also renewed debate about AI and its potential impact on a distinctly human realm such as art. It has broad implications across multiple creative endeavours, including, yes, marketing and advertising.
It’s important to note that AI played a minor, albeit crucial, role in the creation of “Now and Then,” used only to separate and clean up Lennon’s vocals from a muddy-sounding demo tape he made prior to his death in 1980. According to a story in The New York Times, the song features no artificially created sounds.
AI’s use in the creation of “Now and Then” makes a viable case for the argument that it can be a useful tool in the creative process—not as a replacement for human capabilities, but as something that can supplement and augment them.
But after the song’s arrival, director Peter Jackson, who also used the MAL software in his documentary Get Back, speculated that it’s “certainly conceivable” that AI could lead to future Beatles music. And the possibility for AI-generated music extends far beyond just one band. Earlier this year, for example, Music Business Worldwide reported that the AI music app Boomy had created 14.4 million songs.
This is the potentially slippery slope we find ourselves on with AI, particularly as it is only going to keep getting better all the time (it can’t get no worse). We’ve already we’ve seen its ability to faithfully replicate the type of works that have been the sole domain of humans for millennia, but it feels like we’re still at the cave paintings stage of its evolution.
Not surprisingly, governments are already attempting to put AI safeguards in place. Last week, U.S. President Joe Biden signed an executive order mandating safe, ethical, AI development. “To realize the promise of AI and avoid the risks, we need to govern this technology,” said Biden in remarks before signing the order. “There’s no other way around it, in my view. It must be governed.”
But the safety implications are only the tip of the AI spear. Its effect on the workforce, for example, could be profound. To this point, the official line from agency leaders tends to be that AI won’t impact staffing levels, but the things they said today won’t necessarily apply as its true capabilities become more evident.
And in the end, if AI can play a role in the creation of a new song by the greatest band the world has ever known, more than 40 years after the death of its creator, what will it be capable of achieving when you or I are 64?