Why AI-generated Esperanto music?

Why AI-generated Esperanto music?

 

Open AI 500
 

Nowadays we are all confronted with a new phenomenon with the arrival of AI (Artificial Intelligence) for some time now.

ART is also affected in all its fields: graphic, musical, video, cinematic, writing, etc.

The phenomenon also affects Esperanto culture with the appearance of various “works of art” generated by AI.

The purpose of this article is to uncover the phenomenon and have different points of view and opinions so that artists and readers have enough information and arguments to debate the topic, especially here in relation to Esperanto music.

Many questions, doubts, and fears arise about the topic almost everywhere in the world.

Here in France, the FELIN (Federation of Independent Labels) is raising serious warnings about possible damage to musical creation by artists (musicians, composers, authors), and to copyright.

Concerns are mounting after Spotify executives said the average user of the platform would not notice if songs were produced by fake artists. In her book Mood Machine, Liz Pelly reveals “a secret program introduced by the streaming platform Spotify to fill some ultra-popular playlists with songs created by fictitious authors, thereby reducing the share of revenue paid to artists”. Read the article in Le Monde  and an excerpt in Les Jours .

Almost simultaneously, “Deezer opposes the use of AI in music to better pay artists" – an  article in Libération  – « One in ten songs delivered to the Deezer platform is not the work of an artist, but was created artificially, warns the music platform in a press release published on January 24, 2025. To address this, the French company, created in 2007, announced the installation of an “effective tool” to detect the approximately 10,000 pieces of music generated entirely by AI that are delivered to the platform every day ».

In response to this open battle on streaming platforms, 38 international organizations representing the creative and cultural sectors have issued a call to build a future that will reconcile the development of AI with respect for copyright and neighboring rights (music rights in videos, films, advertising, etc.).

From an ecological point of view, considering that a simple search via AI consumes 10 times more energy than the same search via Google, imagine the energy required to create music, a song, a graphic, or a video.

Is it possible to say that a person who uses AI to generate music is an artist?

It is better to say that this person is an “Orderer” because he orders the machine to make and perform the music and song.

Le CHAT (France)

comment utiliser le chat de mistral ai 500
 

Explanations about the process of creating AI-generated music, according to an interview with an Esperanto information scientist from Hungary.

Németh S. Csaba completed his studies in Esperanto and computer science in Szombathely (Hungary) under the guidance of Professor Smidéliusz Katalin. After university, he almost immediately stopped all his connections with the Esperanto movement, only literature remained as a passion. Then came the revolution of artificial intelligence, and everything changed.

Csaba publishes music he created with AI on his YouTube channel.

His goal is to create songs with a modern sound that he enjoys listening to. That is the only goal. But then some songs turned out so “damned good” that he simply felt a strong urge to publish them. When he likes a song, he publishes it for the Esperanto community – but others can also listen to it. Csaba thinks that too few songs appear in Esperanto, so according to him every drop counts.

Like many people, Csaba usually listens to songs that he understands: Hungarian, English, German, Swedish, Russian, Italian, Spanish, Ukrainian, and of course Esperanto.
If he likes a song, for example “Crazy” by Seal, he tells ChatGPT or Deepseek AI his feelings, and asks it to write a song inspired by it. 
He asks the AI to characterize the style of the song and add musical notes to the lyrics.

 

CHATGPT500

ChatGPT is a conversational model created by OpenAI that can understand and generate human language. 
It is trained on a huge amount of text from the internet, and can answer questions, write texts, help with programming, and perform many other tasks. 
The latest version is called GPT-4.

 

Deepseek500

DeepSeek is another language model, created by a Chinese company. 
It is similar to ChatGPT, but it has some unique capabilities, especially when it comes to programming and scientific analysis. 
The model is trained on a variety of sources and can help with many technical tasks.

Then he checks the linguistic correctness in Esperanto, and gives the AI-generated instructions to SUNO.

Suno500

SUNO  is an artificial intelligence specialized in musical creation. It can generate original songs, compositions, and musical arrangements based on the user’s descriptions. The system understands musical styles, and can create various genres.

Unfortunately, he has not yet managed to influence how SUNO pronounces Esperanto words, so a lot depends on luck. However, constant progress can be noted. SUNO generates two versions of the song. He listens to both, and if he doesn’t like either one, he asks for two more, and so on. If a version pleases Csaba enough, he uses SORA to create 10-second video clips for the music.

sora 500
 

SORA is a new video generation model from OpenAI that can create realistic videos based on text descriptions. It can generate complex scenes with people, animals, and objects in movement. The quality of the generated videos is remarkably high and natural.

Because SORA sometimes produces astonishingly confusing things, it is not enough to put together the 10-second fragments – he also has to remove the parts that are “too strange”. Finally, he combines the finished video and sound and shares it either on his Facebook page or on his YouTube channel.

According to Csaba, all these tools represent important advances in artificial intelligence, and show how quickly the technology is developing. They are part of a larger trend towards more capable and versatile AI systems.

He does not believe that real musicians are in danger, because an essential part of music appreciation is also the admiration for the artists themselves. And if such types of music do not exist at all, then there is actually no point in talking about it in the Esperanto sphere.

He does not think that AI-made music will harm or kill originality. However, in a short time, robots will appear in the developed world that will even be able to write their own music on request. For example, anyone can ask them to sing about their greatness, or to create pleasant music for a candlelit date. With the development of video makers, the time will soon come when it will be possible to stop a movie you are watching because you don’t like how the story is developing, and with a simple detailed explanation of how it should continue (for example, that the protagonist’s wife doesn’t die), and the AI will find such a continuation of the story with video, sound, etc. The same is already possible with music. SUNO, for example, is able to continue a piece of music in such a way that it better meets the expectations of the customer....

Here are examples of AI-created Esperanto music videos by Csaba.

 

Climate theme


Esperanto Metal pop

Csaba’s YouTube channel Here!

 

and here is the album Ludovika Poezio created by the AI music
of Nova Tonaro on his Youtube channel
 

Nova Tonaro Ludovika Poezio 500

 

Viewpoint of Johano Ĝo , creator of MV (The social network in Esperanto)

Johano Ĝo250
 

AI songs are fun, and sometimes even good quality, but I prefer that Vinilkosmo avoid AI music.

Soon you won’t be able to differentiate between AI music and real music. If the services continue to remain cheap, we will find more and more AI music everywhere, also in Esperanto.

I’m not against it, but only if you are honest about what AI music is and what it is not.

That is the reason for establishing the group “Esperanto per AI en MV” especially for AI music

One of my favorite music groups from my youth is Daft Punk. For many years, I didn’t know that the group mainly used music from other musicians and rearranged it. My favorite album by them, Discovery, is almost completely devoid of original music. Even while I still like the music, I never really liked that way of working. Most likely because the fact that they didn’t create original music wasn’t clearly explained to me at the time. 
Daft Punk didn’t hide it, but they didn’t reveal it either.

I have a similar feeling about AI music. If it’s clearly shown that something is by AI, I don’t have a problem. However, that’s not the case today.
There’s already a lot of music on Spotify that’s by AI, but the uploaders pretend to be the creators.

I can’t accept deception.

 

Viewpoint of JoMo (Occitania)

JomO cheboksaro koncerto500
 

I have no idea what to think about using AI to compose music. AI has just arrived in our lives, and I can’t guess at all that it will do in the future, what its use will be, and whether it will help humanity, or vice versa. I’m usually optimistic, but this is too important and new for us to be able to predict what the future will look like. Already in some companies, AI organizes the work of projects, organizes meetings between workers, decides who will do what, checks whether the work has already been done.

I have already heard songs made by AI, especially in Esperanto. The music is quite good but not extraordinary, and it is easy to understand that the lyrics were made by it, because they very normally describe a situation and never add a single man-made invention, which gives the charm of a good text. So far, AI does not know how to joke, mock, or imagine.

I can only ask, in the current situation, that it be clearly indicated when a song was made by AI and not by composers and writers. 
Will they disappear?
I doubt it, but I am not sure that it will not.

Only the future will tell.

 

Viewpoint by Mikel Klav (Brazil)

Mikel Klav500
 

Replacing AI for human ability and musical ability is a big reason not to trust AI.
Laws must be passed to regulate the use of AI. 
AI must be used for assistance but not for replacement, and even rules for the platforms must exist because the artistic work must be preserved and protected. Machines cannot surpass the human mind that precisely created the artificial intelligence.

That could even be a musical theme!

 

Viewpoint by Martin Wiese (Sweden)
 

Reflecting on AI music, I want to answer: I hate and abhor it! And now I will try to explain why. It is a two-sided issue: firstly, the process by which music is created can be discussed, secondly, the result of the process – the music itself.

I want to clarify right away that I am not an enemy of new technology, I myself benefit greatly from technical developments when creating my own music. Although my music is almost entirely played by real people with real instruments and voices, I mainly use modern computer technology during the mixing process with many applications, some of which even partially operate with artificial intelligence to analyze and manipulate sounds.

But using artificial intelligence to create a piece of music from start to finish is another matter. For me, music is communication between souls. The musician/composer expresses himself through the choice of chords, the melody, the text, through the choice of various instruments and their sounds, through the timbre and intensity of the voice. Even the imperfections of the performance are part of the expression that is transmitted to the potential listeners, who in turn will decide whether they like it or not, whether they feel an emotional connection or not. But AI music does not come from the soul. It does not want anything. Its creation does not include personal preferences, or the playing styles that come from personal skill and limitations, nor the moments of happiness when the artist suddenly finds the right melody, chord, word, or sound to elevate the piece of music to a higher level. AI music is empty, even if it sounds “perfect”.

But does it really sound perfect? Its development is going very fast, and what I say today about its characteristics may not be valid tomorrow. But I can only discuss it as it sounds today. And so far it is immediately obvious that the artificial intelligence is not aware of what it is doing. Although it can create a sound that superficially resembles music, it has not yet mastered the craft of creating music with a real song structure: verses, stanzas, choruses, side themes, etc. It still mispronounces, and AI-written lyrics are still empty, banal mush. Of course, that will probably change quickly, and in a year or two even I might be fooled into thinking that an AI-made piece of music is performed by real musicians. But that doesn’t mean it will be of equal value to a human-made piece of music, just as an AI image will never be equal to a photograph or a painting. In my opinion, it’s ridiculous when people consider themselves artists just because they’ve ordered an artificial intelligence robot to make music or an image with specific characteristics, just as ordering a painting from a painter doesn’t make the customer the artist.

But why are so many people so enthusiastic about AI art and music? Part of it is certainly because the technique allows people who don’t master the craft to feel like they’re creating something, and also because the result often sounds and looks “professional” in a superficial sense. The sound image is similar to that of world-famous hits, extremely polished, and when there is a voice, it always hits the exact notes with perfect timbre and perfect precision. But much of the expression and personality can be heard in the small imperfections, and since artificial intelligence has no personality, it cannot express it either. Unfortunately, this is not just a problem with AI music, but in the music industry in general. Nowadays, it is standard to always use pitch correction software even for the best singers in the world, because not even the most outstanding singer always sings exactly the theoretically correct tone. By using such software, you effectively steal the personality and expression of the voices and make them resemble artificial voices. This means that music made by real people and AI music are converging on each other, and soon it will not be possible to distinguish between a robot and the manipulated sound of real people. For me personally, this makes contemporary music more and more uninteresting, and a robot voice as well as a “pitch corrected” human voice triggers shivers of horror in me.

Some certainly think that I am against AI music because I, as a musician, feel threatened by its competition. But I always welcome competition from other Esperanto artists who enrich our scene and inspire new projects. I fear that the flood of artificial, “perfect” music will gradually remove from people the ability to appreciate the personal expression of human imperfection. But I am strongly convinced that there will always be people for whom music is more than just sound, for whom, just like for me, music is expressions and emotions from soul to soul. People who want to enjoy live concerts with the direct, immediate communication of real instruments and voices. And people who – like me – make music out of necessity, certainly not always perfect, but sincere.

Martin's hand-drawn music video, drawn with a special iPad pencil in Adobe Fresco, a free iPad app.

 

 

Viewpoint of (Jean-Claude ) Giuzeppe Patalano (Occitania)

It all depends on how AI is used.

Before the advent of computers, artists often used a rhyming dictionary to write songs; this is an organized database, published to help writers.

With the advent of computers, a large number of creative tasks could be taken over by machines. AI can be a good tool if it helps improve the artist’s work.

In their creative work, the artist necessarily uses their knowledge, talent, skills, and the work of their predecessors, their “influences”. Their role as a creator is to bring their personal touch to this experience to create an original work.

The danger is when AI is used to completely replace the artist. This means that the AI has “learned” the artist’s original work, so its production will be an imitation of human work, i.e. plagiarism.

For humans, plagiarism is considered theft; in short, it is like making counterfeit banknotes.

Here is the latest music video from Giuzeppe made with very limited resources and a few free programs.

Article compiled by Flo, in order to open a debate...