What! AI generates Arijit Singh, Atif Aslam, Sidhu Moosewala, Sonu Nigam’s voice: What’s at stake?

Imagine Bedardeya from Tu Jhoothi Main Makkaar, originally sung by Arijit Singh, in the voice of late singer KK. Or Gerua in the voice of Atif Aslam. Or even late Sidhu Moosewala’s voice in Tumhe Dillagi, which was sung by Nusrat Fateh Ali Khan. All this isn’t some fantasy anymore- you can just go on YouTube, and search for hundreds of such covers, which we come across as we are scrolling.

AI (Artificial Intelligence) has been getting a lot of attention of late, with even the writers in the West bringing it up. However, here the line between reality and artificial are getting blurred.

USING A DEAD ARTIST’S VOICE

Shaan, singer and a close friend of late KK, isn’t happy with AI being able to generate KK’s voice, “It’s very insensitive, and complete invasion of that person’s privacy, without them being aware. Making their voice come alive, it is morally wrong somewhere. Today people have patents on their voice, even if alive, they won’t allow you, since that voice belongs to them, it’s their pehchaan. Would they even want to sing the song if alive, or have the same expressions? As fun and spoof it is still okay, but should not become a practice. I would not want my KK’s voice any other way than the way he sung.”

Isn’t it a positive, that the artist can remain immortal if their voice can be generated via AI? Shaan reasons, “A musician is immortal anyway through his/ her own work. It’s not like it’s on tape or CD and will wear off in sometime.”

WHAT DOES THE LAW SAY?

Safir Anand, senior partner, Anand and Anand, recently represented Amitabh Bachchan in a case where his voice was imitated to promote a lottery ad. The Delhi HC barred the use of Bachchan’s voice, image or name without approval. Addressing the use of Ai to recreate a singer’s voice, dead or alive, he says, “Singers have performing rights and moral rights in their works and in the ordinary course of business, do have recourse to remedies under the Copyright provisions in India. Such rights can also be enforced and exercised by the trusts and legal heirs of artists that are no longer alive. Even copyright societies enforce rights on behalf of the creative community. AI generated music and its ability to violate third party rights is as yet an uncharted issue not just in India but globally.”

There is a global debate ongoing on whether works created by AI itself could be considered for copyright protection at all considering that they have no human author’s input, he continues.

AI IS GOOD, IF USED CAREFULLY

Music composer Salim Merchant feels AI could be useful if employed carefully. “If somebody is not accessible, you can use their voice for something, that’s the advantage. But is it fair to the dead artist’s family or heirs, that you are recording a song with a singer through AI, and not paying for it?” Is it a threat to existing artists and their profession? “The one thing Ai cannot do is emote. You need a source of stimulation, someone has to originally song it! In Bedardeya’s case, Arijit sung it. Without a source, how will you create an AI model?,” he makes it clear.

Nakkash Aziz uses AI to generate dummy words while writing lyrics, “There’s nothing wrong as long as it is used ethically and doesn’t hurt anybody’s sentiments. It would be very sad if someone made money out of using KK’s voice.”

Citing AI as a good production tool, Ehsaan Noorani, one part of the Shankar-Ehsaan-Loy trio quips, “I use a little plug in on my system and it helps in creating a music melody, but what it does is not enough. Even if the voice recreation bit is used for fun and as a hobby, it is still a bit dicey, unless you are doing it with the permission of the dead artist’s surviving family/ heirs.”