An earthquake shook the music industry on 18th April. The biggest shake-up since the mp3. It will have huge repercussions.
You may have heard last week - a song purporting to be Drake and The Weeknd was pulled from Spotify and other streamers by Universal Music Group (UMG) after going viral over that weekend. For “infringing content created with generative AI”.
'Heart on My Sleeve' was first posted on TikTok by ‘Ghostwriter977’ (see left) and subsequently streamed. Two days later it was removed. In that short time, it had scored 15m TikTok views, 600,000 Spotify plays and 275,000 YouTube views. Blimey, no wonder UMG were angry. The song featuring Drake’s ‘vocal’ was a breathtakingly convincing fake.
To date, we don’t know who Ghostwriter977 is. There is hilarious speculation this was either the best marketing stunt ever by a tech start up. Or, even Drake himself was in on it, and in fact, it IS him singing. Either way - if true, it would be the smartest, most brilliant marketing strategy.
TikTok and YouTube are awash with songs using AI-generated vocals purporting to be famous artists. Kanye West scores the most fakes. I gave up after counting 30. The technology and sheer quality of the ‘fake’ voice is uncanny. It can be very, very convincing. Even to fans. However the recent ‘Drake Fake’ takes it to another level. And some.
Anyone has access to this web-based technology. Hosted on Google, it’s here. Please check with your IT department before running this code!
UMG said the viral postings “demonstrate why platforms have a fundamental legal and ethical responsibility to prevent the use of their services in ways that harm artists”. Possibly an understatement. It’s a clear case of ‘passing off’. Anyone in advertising music knows you can’t use a singer who impersonates another known living artist without their permission. You’d be royally sued. Somehow those norms don’t apply in TikTok land.
We’ve been experimenting with similar technology here - I should add, not for passing off as well-known artists. We use it for working out harmonies, blocking out backing vocals, and creative experimentation. Usually, before a human singer does it for real. A human vocalist will take considerably less time.
Brands and advertisers wanting to use AI fake covers as part of their music strategy should exercise caution for now! Unless perhaps you’re an up-start disruptor company and don’t mind the legal flack.
However, AI-generated voices are just a precursor to a bigger storm.
Aside from the voice, AI can ‘learn’ from the actual music of any composer you like, who has a distinctive body of work. Beethoven, Abba, Drake… Dua Lipa… Mr Blobby… Using their catalogue of songs, AI can compose very, very convincingly ‘in the style of’ the artist of your choice. Both Beethoven and Schubert’s ‘Unfinished’ Symphonies have been ‘finished’ with the help of AI.
The FT said it had become aware that certain services had been trained on copyrighted music 'without obtaining the required consents'.
Last year the Recording Industry of America said AI companies were violating copyrights en masse by using music to train their machines..
In response to the Drake Fake, UMG said, “The training of generative AI using our artists’ music (which represents both a breach of our agreements and a violation of copyright law) as well as the availability of infringing content created with generative AI on DSPs [digital service providers], begs the question as to which side of history all stakeholders in the music ecosystem want to be on: the side of artists, fans and human creative expression, or on the side of deep fakes, fraud and denying artists their due compensation.
Toby Jarvis - fonder and creative director @ A-MNEMONIC Music
I can’t help wondering if one of our writers (all of whom have a distinctive original sound) were to discover AI was impersonating their writing, how enraging that would be. Perhaps if I fed all my own music (a not insubstantial body of work), could I retire to the Caribbean and let my algorithm do all the heavy work? It wouldn’t be as much fun.
The established music biz has always been viewed with contempt by young people. Perhaps healthily so. And, as we know there’s no better way to sell a record than to ban it. Well, this is GenTikTok taking control. They have savvy, humour and creativity.
And access to mind-blowing tech.
I’ve always known authenticity to win out. However, guessing where this inherently in-authentic trend will go is tricky. Maybe some form of licensing between the rights holders and AI companies? Hopefully, authenticity and discerning audiences will win through.
Grimes
Singer Grimes had other ideas. “Anyone can use my voice for AI-generated songs - I’ll split 50% royalties on any successful AI-generated song that uses my voice,” … “Feel free to use my voice without penalty,” she said, claiming she has no label and 'no legal bindings.'
She may live to regret that.