Back in November 2023, Google's artificial lab in London announced a generative-AI (artificial intelligence) music technology. Titled Dream Track, it was an experiment to produce a unique soundtrack with the AI-generated voice and musical style of several artists.
The artists included Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Sia, T-Pain, Troye Sivan, and Papoose. But the thing was, each of these artists were officially partnering with Google to 'help it test and learn to shape the future of AI in music.'
Google also created a tool called SynthID, which can watermark and detect synthetically generated content.
Google's deep dive into the ocean of AI-generated music had distinguishable safety protocols and guardrails to ensure copyright and ethical validity, as the world of music did not have a pleasant experience with its first major public use of AI in mainstream music.
In April 2023, a song featuring the voices of Drake and The Weeknd called 'Heart on My Sleeve' amassed over 250,000 Spotify streams and 10 million views on TikTok.
Just that the song was an upload from an anonymous source named Ghostwriter, who had put AI to use to recreate the voices of Drake and The Weeknd. Not stopping with that, Ghostwriter subsequently released a new track, 'Whiplash,' that mimicked Travis Scott and 21 Savage.
Though the songs were soon taken off from streaming platforms, the world of music was caught staggering as a new Pandora box was opened.
In India, TikTok is banned and hence some of these developments haven't caught the headlines as they otherwise would have. For, on TikTok, there is a spate of AI covers of musicians who have not actually sung the song.
Every kind of twisted experiment has been attempted. From re-generating dead singers' voices to making existing singers sing in a style that is not theirs. Like Kanye doing Coldplay or Queen. And there is also AIsis, an AI cover band of Oasis, which has already come out with a 32-minute album, The Lost Tapes.
Dead Singers Bamba Bakya And Shahul Hameed Get Their Voices
As it happens, this month, a big AI-generated musical controversy has hit closer home.
As the man who is seen as a pioneer in India in the use of technology in film music, it is A R Rahman who is inevitably in the centre of this maelstrom. He has put to use the voices of Bamba Bakya and Shahul Hameed — two singers from Tamil Nadu who have passed away — for the song Thimiri Yezhuda for Rajnikanth's new movie Lal Salaam (scheduled for release on 9 February).
The song itself is just about middling. The voices in themselves, like most of modern-day singing, don’t have much character. These are besides the point though.
The recording label Sony, said that the song was recorded using the tech of 'Timeless Voices', which is a company that is said to 'preserve voices of legendary artists using the power of artificial intelligence'.
After a quiet release, a small controversy erupted over the various issues surrounding the recreation of the voices of dead singers through invasive technology.
On social media platforms, where everything is a raging issue, this understandably blew up. And responding to them, Rahman posted: "We took permission from their families and sent deserving remuneration for using their voice algorithms ..technology is not a threat and a nuisance if we use it right…#respect #nostalgia."
But Rahman's deadpan response hardly answers the plethora of issues surrounding the use of AI in the musical industry.
While taking the consent and compensating the family of the deceased artists for use of their voices seems a fair thing to do, the larger question to ask is: Does the family own the voice? Is it a property to be legally legacy-ied? Is a person's voice copyrightable?
Imagine Kesariya In Kishore Kumar’s Singing
In many countries including the US, artists own their voices and likenesses, which are protected not by copyright but by the 'right of publicity'.
In India, things in this field are very hazy. But experts feel the unlicensed use of AI-generated voices and styles can be a form of identity theft. Even permitted use still leaves room for moral claims. Whether copyrighted material can be used as training data for AI is another area of limbo.
So it is both a legal and ethical conundrum. And make no mistake about it, there are no easy answers. Like with deep fake videos, which are creating havoc, AI-generated music too comes with its own torrent of problems.
To be sure, disruption due to technology in the music world is not new. But AI-generated music is not a format change in the way music is consumed, like the transition from records and cassettes to CDs. It is also not about using programmable techno instruments to create music. It is also not just about business and copyrights.
AI cuts much deeper as it hits at the very core of music — creativity and individuality. Human experience is also at stake here. Many of us sometimes imagine what it would be like for Kishore Kumar or Mohammad Rafi to attempt, say, the Kesariya from Brahmastra. That nostalgic imagination in our heads is now a scientific reality that blunts our artistic sensibilities.
The copyright and monetisation part will take several years before any sort of clarity can be arrived at. The fundamental issue is: who owns the AI-created music, the person who used the AI or the artist whose style is the inspiration?
As things stand now, the copyright is vested with the former. In other words, he or she is the owner of the music that they didn't actually create.
The Genie Can’t Be Put Back In The Bottle
When a company, the size of Google, is making investments in developing products and tools for AI music, things indeed seem dire. The music industry itself is worried about the developments.
According to a PRS survey in August 2023 (PRS is a British music copyright collective), 74 per cent of its members are concerned about AI-generated music competing with human-made compositions.
PRS said: “93 per cent believe creators deserve to be compensated if their music is used for AI-generated content, and 89 per cent feel that AI tools should be transparent about how they generate AI works."
As many of these AI musical attempts are from anonymous sources, the music industry is now more focused on crippling the distribution system.
Music companies are targeting major streaming services like Spotify and Apple requesting that they block AI companies from using AI music of their artists. The music companies are also issuing a plethora of take-down notices to various platforms. But enforcement has been a problem in many countries. And it'll be so in India too.
All said and done, the genie is out of the bottle and can't be put back in. But let's be clear: AI is not a creator in the traditional sense. It can be fun for Instagram reels and evanescent use. There is no real innovation or skill.
AI just synthesises an imitation of human content, which is derived from algorithms that study it. Hence the use of AI seems to be a lazy way to create music. Ironically, Rahman himself said as much in his recent talk at Oxford. For him, to attempt something that he himself avowedly seemed not enamoured of, is a bit bizarre.
No wonder, he is facing the music now.
An appeal from Swarajya
At Swarajya, we rely on our readers' support through subscriptions to sustain our media platform. Unlike larger conglomerates, we are unable to relentlessly chase advertising money — our model is largely built on your patronage.
Your support has never been more crucial. We work tirelessly to deliver 10-15 high-quality articles daily, ensuring you receive insightful content from 7 AM to 10 PM.
If you believe India's story has to be articulated in a way it has never been done before without shrugging it off, become a patron (or) subscribe now for ₹̶2̶4̶0̶0̶ ₹1999 and get 12 print issues, unlimited digital access for 1 year, a special India that is Bharat T-shirt (Offer ends soon).
We are counting on you!