Streaming may have killed off MP3s in the same way that the iPod killed off CDs and cassettes ended the dominance of vinyl.
Each time the music industry has adapted and evolved.
But what do record labels have up their sleeves to protect themselves from the threat of artificial intelligence (AI)?
Well, after a spate of ‘deepfake’ songs convincingly impersonating the likes of Johnny Cash, Frank Sinatra and Drake, the world’s biggest record label is taking action.
Universal Music is now in talks with Google to license artists’ voices and melodies so they can be used in AI-generated songs. So how will this new age music be produced and what does it sound like? MailOnline take a look.
Shake up: Universal Music is now in talks with Google to license artists’ voices and melodies so they can be used in AI-generated songs. It follows the rise of a series of ‘deepfake’ songs where the voices of Drake and Rihanna have been convincingly imitated.
Convincing: Harry Styles’ voice has also been mimicked to create a duet with American singer-songwriter Lizzy McAlpine (right), whose song ‘Ceilings’ went viral on TikTok
How do AI songs sound?
There are a few examples, namely a cloned Frank Sinatra singing Coolio’s Gangsta’s Paradise, ‘Johnny Cash’ doing a Barbie Girl cover of Aqua and Harry Styles’s voice imitated to create a duet with American singer-songwriter Lizzy McAlpine, whose song Ceilings went viral on TikTok. .
In fact, some YouTube channels are even dedicated to creating AI-generated music.
Kanye West’s voice was taken to sound like he’s singing the 2006 acoustic ballad Hey There, Delilah, while a ‘deepfake’ of Rihanna supposedly performing Beyoncé’s Cuff It was produced.
Drake and The Weeknd have also fallen out of favor with the growing trend, while technology has helped ‘revive’ the voices of dead musicians like Elvis Presley, David Bowie and Michael Jackson, as well as Cash and Sinatra.
A YouTube user named PluggingAI also promotes songs that imitate the voices of dead rappers Notorious BIG and Tupac.
False: there has been a ‘deepfake’ of Rihanna supposedly performing Cuff It by Beyoncé
Mimicked: Kanye West’s voice has been taken to sound like he’s singing the acoustic ballad Hey There, Delilah
How is AI-generated music created?
Essentially anyone can do it if they know how.
Several websites already offer fans the ability to create new songs using similar voices belonging to some of the biggest stars in the world of pop.
One, created by the California-based company OpenAI, which is responsible for the hugely popular ChatGPT artificial intelligence bot, is called Jukebox.
It’s a neural network that generates unsettling approximations of pop songs in the style of multiple artists.
Some experts believe the technology could shake up the music industry by creating new hits, but controversy surrounds it due to copyright concerns.
Deepfake music blurs the line between using a copyrighted song and using a cheaper or royalty-free approach.
Is Elvis back from the dead? The Jukebox online library has the potential to create royalty-free knockoffs or even new tunes that could be released as singles under the direction of a deceased artist.
“If someone hasn’t used the actual recording, you wouldn’t have any legal action against them in terms of copyright in respect of the sound recording,” said Rupert Skellett, legal director of British record label Beggars Group.
Another freely accessible AI tool is MidJourney, which generates videos, while French DJ David Guetta has used uberduck.ai to imitate Eminem’s voice so he can add it to one of his instruments.
“I’m sure the future of music is in AI,” Guetta later. he told the bbc.
Are Universal and Google close to producing music?
No, discussions are at a very early stage and no product launch is imminent.
However, the goal is to reach an agreement whereby more software can be developed that allows fans to create songs and pay copyright holders.
Artists would have the option of whether or not to participate in the venture.
It’s a similar situation to what the music business found itself with the rise of YouTube, where people started using hit songs as soundtracks to videos they had created.
This led to years of legal wrangling over copyright infringement until a deal was reached in which the music industry pays around $2bn (£1.5bn) a year for clips generated. by the users.
What do the artists think?
Some artists have complained about the trend. Drake responded that an AI-generated song featuring his vocals was the “last straw.”
His record label also recently fought to get a viral AI song that used his likeness removed from the internet.
Star power: Beyonce’s song Cuff It was ‘covered’ with the help of an AI version of Rihanna’s voice
The Plain White T’s had a hit with Hey There, Delilah in 2006. A cloned voice of Kanye West was used to create an AI-generated version of the song.
In April, Universal Music Group pulled an AI-generated song featuring the simulated voices of Drake and The Weeknd from streaming services for “infringing content created with generative AI.”
The track went viral, and by the time it was removed, it had been played 600,000 times on Spotify and received 15 million views on TikTok and 275,000 on YouTube.
The song, Heart On My Sleeve, is believed to have been created using artificial intelligence programming trained on the artists’ music, something UMG said “represents both a breach of our agreements and a violation of copyright law.” .
However, not all musicians are against AI.
Canadian singer Grimes told the artists that they can use your voice on AI-generated songs for a fair split of 50 percent of the royalties.
The 35-year-old singer, whose real name is Claire Boucher, said nothing was off limits and her “ultimate goal has always been to push the limits rather than have a good song.”
How does the music business make money?
It remains to be seen, experts say. A lot will depend on how popular the AI music turns out to be.
But if a deal can be reached, then it would allow royalties to be paid to record labels, meaning the music industry would not lose out on significant profits if the niche market takes off.
The AI also cloned The Weeknd’s voice and Drake said a song imitating his voice was the ‘last straw’.
Universal certainly believes that a deal with Google is the best way to thwart one of the biggest threats to the music business.
Michael Nash, Universal Music’s executive vice president and chief digital officer, wrote earlier this year that the rise of AI could be a “calamity” for the industry.
He added: “These developments have raised deep concerns in our industry, with similarities drawing between the rise of AI and the rise of Napster and license-free music sharing more than 20 years ago.”
Universal Music General Counsel Jeffrey Harleston also highlighted industry concerns about the technology when he told US lawmakers last month: “An artist’s voice is often the most valuable part of their livelihood and image. public, and steal it, no matter what the means. , That’s wrong.’
Google, which owns YouTube, has also been in talks with Warner Music about a product, the Financial Times. reported.
For its part, the search engine giant would benefit from creating a music product because it would allow it to compete with rivals like Microsoft.
The Bill Gates-founded software company secured a £10bn deal with OpenAI, giving it a head start on AI tech bets.