SummaryWhat is the future of lyrical and musical creativity with the growing AI market in this field? Thousands of music tracks are produced in the blink of an eye, and generated works now account for over 40% of new productions on streaming platforms. Notably, many of these works meet high quality standards and deceive listeners, who do not easily discover they are technically generated. Is there a way to preserve the authenticity of this art, or is it inevitable to agree on a means to benefit from these technologies without eliminating the human element?

The hostile discourse against AI-generated music seems to be receding globally in a notable way. A year ago, there was absolute rejection of this approach, and its opponents held the view that personal imprint cannot be duplicated and human talent cannot be imitated, noting that music and lyrics born from a poet's soul and the human voice will remain the listener's first choice. But the surprise is that, according to the latest reports, listeners themselves can no longer tell the difference between a real and a virtual singer, or between music made entirely by humans and those produced by algorithms. In fact, thousands of such works are published daily among audiences, achieving millions of views and consecutive likes.

It is established that any creative productions through AI applications are fundamentally based on previous human works, synthesized and inspired by them. However, the volume of output, estimated by the famous French streaming platform Deezer at two million tracks per month, suggests that this market will dominate and become mainstream, with human works completely disappearing in the future. Consequently, new measures are moving towards finding a way to deal with this digital music flood, not by rejecting it entirely, but by attempting to regulate it, establishing mechanisms and controls for its use to ensure it does not deceive listeners, while also preserving the space for personal creativity, protecting creators' rights, and warding off the specter of mediocrity from the industry.

But the complexities facing this reality are increasing. At a time when companies have signed multi-million dollar contracts with virtual singers after they entered the top listening charts, other companies oppose this trend and accuse the generators of violating intellectual property rights, as they derive their works from well-known classic songs with which they were fed. There is a third group that insists that AI in music is a fait accompli that cannot be escaped, and they have indeed begun presenting works that blend human elements and algorithms.

Digital Singers

A few weeks ago, a song by Egyptian singer Sherine Abdel-Wahab spread and garnered public admiration, before it was revealed that it was merely a simulation of the Egyptian star's voice, whose fans are eager to hear new material, and that the song was entirely AI-generated. Prior to that, statements by Arab pop stars rejecting the theft of their voices in this way had been pouring in, indicating their opposition to AI methods in the singing field in general. But it seems that innovators and developers are no longer just imitating the vocal imprints of celebrities; instead, they are creating a complete artistic experience.

These singers appear in concerts and music videos, have their own popular songs, and their listeners only discover by chance that they are just virtual people digitally generated. Among them are Lina Nour, Lara Hadid, and Lara Aredo, who combine beautiful voices, looks, and singing in various dialects skillfully. In fact, some of these digital characters also appear in virtual programs and talk about their experience in the field. Likewise, the world was buzzing with songs by a new band called 'The Velvet Sun Down', which achieved millions of listeners on platforms including Spotify, but it later turned out that the band was created using AI and does not exist in reality.

Artificial intelligence has become a reality in the music world (social media)

As for the singer 'Nine', she received offers from major producers and prominent distributors in the artistic field to collaborate with her, as they all believed she was a real singer with a strong voice and that her song choices were popular. All this happened during the first three months of her appearance on social media. Afterwards, her creator, composer, singer and distributor Mohamed Shafiq, revealed that she is a virtual character. Shafiq is now celebrating the first advertisement starring the beloved singer for a soft drink company, noting that her schedule is full of similar offers from brands in fashion and beauty.

The musician, who has a track record of successful works with stars such as Tamer Hosny, Tamer Ashour, Ramy Gamal, Mohamed Ramadan, and Ramy Ayach, explains: 'I studied at the Higher Institute of Arabic Music, and also completed my studies abroad. Currently, I am working on my master's project in artificial intelligence, in addition to my long professional experience. Therefore, I fulfilled all legal and intellectual property requirements before creating this project. She is like a singer I adopted to ensure she would respond to my artistic vision to achieve success, unlike what happens in reality. Many female singers, especially newcomers, do not want to take artistic risks and prefer to play it safe with repetitive content. Also, at the first big financial offer, they leave the companies they started with. So I created a virtual singer with a growing audience and her artistic experiments are in the spotlight because I handle everything about her.'

Even specialists may fall into the trap

The platform Deezer, which has about 20 million monthly active users, nearly half of whom subscribe to the paid version, has launched a free tool to detect the authenticity of songs. It works with about 20 popular streaming platforms, and through it, one can analyze the nature of each work and the extent to which its creator resorted to algorithms, whether fully or partially generated. Deezer, which is very active in this field, indicates that more than 40% of music creators with active accounts on the platform publish AI-generated playlists.

Despite this, music producer Mohamed Shafiq doubts the existence of any mechanism currently to distinguish between AI-produced and traditional songs. He adds: 'There is no mechanism that can detect whether a song is AI-produced or not. Apart from music videos where AI is used obviously, I legally purchase song lyrics, compose the melody and arrangement, in which I also use AI. This is not called cheating, especially since I have already announced it. Moreover, the human element is strongly involved in the artistic process. The result is a good song completed in a short time with a strong voice that has an audience, and I make a profit without getting into human squabbles and disputes. If I hadn't announced it, no one would have discovered it.'

Virtual female singers are more attractive to producers seeking investment on a small budget (social media)