AI creates complete music in any style and language – 03/21/2024 – Ronaldo Lemos

Everything we know about music can change quickly. On March 21st, version 3 of a service that creates music using artificial intelligence called Suno was launched. The service is shocking. It could produce a seismic shock in the music market. With a few lines of text it is possible to create a complete song, in any style, sung in Portuguese, English or other languages.

The songs have “radio quality” (the company’s words), including the arrangements, instrumental bases and pronunciation. I experimented with different styles such as K-pop, MPB, punk, jangle, country music, Brazilian rock and the results were surprising for everyone. I made birthday songs for family members, jingles for our research institute, and even a goth rock song about the dangers of artificial intelligence.

Listen to a song created in Suno here.

It’s even difficult to say everything that’s happening. I will select some points. The first is copyright. It is not difficult to imagine that many people will reduce their consumption of recorded music to listen to their own music created by AI. For example, restaurants and gyms will be able to generate their own music with artificial intelligence. Films and commercials will also have scores made by AI. As there is no authorship or registration, there would be no copyright collection.

The second point is customization. Songs can be created talking about themes familiar to each person. It is possible to create a ballad for your girlfriend, including her name. Make a punk rock talk about your group of friends at school. And so on. The songs will be of good quality and the lyrics will talk about things that matter to each person or group. This is very powerful. A DJ can create a personalized playlist of songs for a wedding party, all of which speak of the bride and groom’s love.

The third point is authorship. Who makes these songs? I sent the songs I generated to people saying “look at the song I made for you”. But did I do it myself? My job was to type a few lines of text and the music appeared thanks to the AI. We are entering an era where creation will become the art of writing precise text that can be interpreted by the machine. Will this be enough to generate copyright? Today this question has no answer.

The fourth point is power. Using the Suno tool, anyone has at their fingertips the power of everything that has ever been created in the history of music. It is possible to invoke any type of composition. This, of course, can become a weapon. For example, in the 2024 election campaigns there will be jingles made in this way, capable of going viral and changing the way we think about political communication.

The fifth point is the flood. Open platforms like YouTube and social networks will be flooded with music made by AI. The price today to make 500 songs per month is US$8 (about R$40). In a year, for the same price, there may no longer be a limit. Millions of synthetic songs could quickly occupy the space of recorded music.

As Jacques Attali said: “sound is prophetic, it is what comes first”. These changes are, once again, the spearhead for us to see the greater challenges we will face from now on.

Gone – recorded music

It’s time to ask: who has the copyright to AI-generated music?

Coming soon – ask what content the Suno service was trained with (the company does not reveal)

LINK PRESENT: Did you like this text? Subscribers can access five free accesses from any link per day. Just click the blue F below.