“This is the last straw”: Your next favorite song is made by Artificial Intelligence

Next song Can be created with a favorite AI. In recent months, new computer programs have been developed that can learn and copy any artist’s vocal modulation and intonation. Thus, users can have the ‘fake’ sound produced by the app cover popular songs by other artists or sing entirely new lyrics.

This possibility has led to a trend of experimentation that has become popular on platforms such as: TikTokwhere did it go viral Rihanna sing the ‘handcuff’ song related to beyonce; Ariana Grande “Passionfruit” by Drake; and ‘Love Yourself’, by Justin bieberreinterpreted by Kanye West. Creating songs with the voices of other artists has never been easier.

Frederic Font, a researcher in the Music Technology group at Pompeu Fabra University (UPF), warns that the potential is great and there is “still something very experimental” when it comes to music.

The expert, who will lead a talk on the topic at Sónar +D (which will be the main axis of his program with the influence of artificial intelligence in art this year), said, “There is a part of artificial intelligence research where the aim is not to completely change the creation process, but to help an artist think differently or to help an artist think differently. to contribute creatively, to provide them with new possibilities”. “If you look at it from that perspective, it’s not so groundbreaking, that’s what’s happened throughout the history of music as new elements emerge,” he adds.

record label scare

However, the fact that all this content is produced without the consent of those whose audio technique is stolen worries both. artists like grown-ups record tags representing them who are aware of this change Universal Music Group already asked the major streaming music platforms – among them spotify And Apple Music— Don’t let AI companies gain unauthorized access to your music catalog to train their programs. “We will not hesitate to take action to defend the rights of ourselves and our artists,” they threatened in an email the Financial Times has access to.

But, Font says, “you can’t train with as much data in music as you can in other areas like images or text, and then the quality won’t be as good.” “Right now we’re dazzled, but we need to see the real impact this can have over the long term,” says the UPF researcher, and in any case it “will also depend on the weight society puts on it.” on technological possibilities”.

Sony Music, which this newspaper consults, does not want to evaluate the situation and difficulties presented by artificial intelligence for now.

From Drake to ‘resurrection’ Oasis

Last Sunday, a user TikTok nickname ‘ghost writer977He released “Heart on my sleeve”, a song created with AI to reproduce the voices of Drake and The Weeknd. After more than seven million views, the song made the leap to Spotify and Apple Music, where it was listened to by more than 600,000 people in less than 24 hours. But on Tuesday the song is no longer available on any of these three platforms or YouTubealthough it finds other ways to stay roaming.

It is unknown whether the record companies want to remove that song and what intentions are behind this mysterious account. Is this a gimmick by Drake and Universal to get our attention, or is it a violation of our rules? Copyright from an opportunist? TikTok, contacted by this newspaper, did not want to give details on the case, but reminded that their policy does not allow “content that infringes on the intellectual property rights of others.”

In a busy week in the AI ​​/ music relationship, we also witnessed the ‘resurrection’ of Oasis thanks to artificial intelligence. In fact, this fake meeting of the Gallagher brothers was christened AISIS. [AI son las siglas en inglés de Artificial Intelligence]. In this music scenario bordering on science fiction, does AI pose a threat to artists and creators? “I really don’t think it could pose a big threat, it could mean that something is transforming. This will ultimately depend on the listener’s attitude towards it. In the case of traditional music consumption, the connection between listener and listener and the artist goes beyond the music itself, the sound file itself, and what these systems do in the end is to create sound files,” Font replies.

Battle for copyright

Universal’s request is part of a general concern after the advent of pseudo-generative AI. fear automation caused by such tools. ChatGPT increasingly affecting more and more professions, leading to a backlash from those who might be affected. Illustrators and graphics companies have already denounced renderers such as: Stability AI, devianart anyone mid trip for infringing copyright. Record companies can follow the same path. “The traditional concept of copyright has already been questioned in the digital age. In musical creation, too, with the problem of exaggerated examples. Finally, these systems are taking a step further,” Font argues. “Perhaps it is the concept of copyright that we have that needs to be questioned. This is another indication that perhaps the framework needs to be rethought in some way,” he adds.

The impact of artificial intelligence on music goes beyond synthetic sound generators. And Google developed a language model, MusicLM, capable of generating music from text descriptions. While the company hasn’t released it due to “risks associated with possible misuse of creative content,” this prospect will pose another challenge for the music industry.

artistic opportunity

the first book to explore the idea computers can help compose music, was written in 1959. As Font pointed out at the beginning of this text, computer evolution is accelerating a possibility that many extremist artists prefer to see as a creative opportunity. This situation boat, holly, bull and moi anyone Ash Kooshabut it also – albeit at a more primitive stage – did David BowieExperimenting with a random expression generator.

In the cases mentioned above, it’s the artists themselves who clone their voices with AI to innovate. The risk arises when there is a third party seizing it. More recently, DJ and producer David Guetta used this technology directly in one to recreate the rapper’s voice Eminem. The complex debate around this issue led the Frenchman to state that he would not market this creation.

Faced with the enormous challenge it faces, the music industry will have to adapt. Some platforms have been doing this for a long time. This is the case of Spotify, which is working on a set of artificial intelligence tools that allow users to compose and remix synthetic music from various artists, as well as Apple Music, which bought a startup specialist in this technology last year. In the UK, authorities are even investigating whether artificial music can have intellectual property rights. What will happen?