In May 2025, there was a club track in Ukrainian on YouTube, “The Mist Dances Across the Earth” («Туман танцює по землі повільно»), by Y.K. Music. Very soon, this song became a hit, reached over 23 mln streams in total with YouTube Music application, and still remains among the top-20 most popular releases in Ukraine. This success story has one noteworthy detail – like all the other artworks by Y.K. Music, this song has been generated by artificial intelligence.
NGL.media decided to analyze how many such songs are present in Ukrainian top charts and whether such content can push the music, created by actual people, away.
How it all started
Oleksii Bondarenko, the editor-in-chief of LiRoom, the media on culture, and the co-founder of an indie rock group, Black Balloon, believes that while AI was used as an instrument, there was no problem.
“There have long been instruments of artificial intelligence to create music. I use them myself for mastering the final stage of processing a sound record, but you will never feel any processing there. And it is ok, because a listener doesn’t care which compressor was used to tune the sound. But when artificial intelligence started creating music completely, that’s when the issues were raised,” Oleksii Bondarenko explains for NGL.media.
Suno is one of the most popular platforms for music creation and processing. At first, musicians were using it to improve the voice, to adjust frequencies and volume. But the possibilities of the platform have expanded so much that now one can create a full-fledged album in one style in several minutes.
“It started when, for example, Spotify used such platforms to create artificial tracks of functional music, played in gyms or restaurants. The motivation was simple: to avoid paying royalties to actual artists, whose music had previously been included in such playlists,” Oleksii Bondarenko tells.
And then there have appeared entire albums, created by AI, with made-up stories about a group that has allegedly made it. For instance, this October, there was a scandal around a seemingly Ukrainian-Lithuanian punk group, Transportna. The musicians have 18 thousand subscribers on YouTube and hundreds of thousands of views and streams. However, after the song “As if in a Movie”, it has become clear that this composition was completely AI-generated. It was caused by the sloppiness of “musicians” – at the end of the track, there is the text “the sound of rain somewhat increases and gradually absorbs the music, remaining alone for a few more seconds, before cutting off”. Instead of following the instructions, the AI just added it as a part of the text.
“Generally, the prevalence of AI-tracks started in late 2024. Last year, I summarized and counted 101 Ukrainian artists, behind whose voices and music there definitely was 100% AI, as of this October, there are over 700,” Oleksii Yefymenko, the founder of the analytical platform NUAM, explained to NGL.media the service to trace and study the Ukrainian music.
According to Yefymenko’s estimates, 34% of all music released in Ukraine this year has been generated by artificial intelligence. By NGL.media’s analysis, as of September 2025, the YouTube top chart contained 14% of AI-generated songs. And, while at the beginning of the year, these were one or two songs out of a hundred, now there are dozens of them.
How did we determine that a composition had been created by AI?
We investigated the spreading of AI-generated music on YouTube Music, which is the most popular streaming music service in Ukraine there are no data for the period of the full-scale invasion; the last available data demonstrate that YouTube Music is twice as popular as Spotify. Apple Music and Deezer cover much smaller shares of the market in Ukraine. Every week, YouTube publishes a rating of the 100 most popular songs in Ukraine. We downloaded all the songs that entered this rating at least once from January to September of 2025 (there is no point in taking the previous years, since the AI-instruments of generating music were not so widespread yet).
A total of 586 songs were analyzed by us using two instruments. The first one is AI Music Detector by the French institute, IRCAM, which is involved in the study of music and sound, and the second one is AI Song Checker by SubmitHub, based on the study of synthetic music from such popular platforms as Suno and Udio. We chose these two instruments after initially testing several compositions that were obviously generated and some popular songs, which were undoubtedly written by humans. Since the AI Song Checker provides more accurate results, we selected it as the basis for our graphics.
In addition, the analysts of the NUAM platform, investigating new releases in Ukrainian music, shared their analysis of AI-generated music with NGL.media.
We presented weekly YouTube ratings since early 2025 in the form of columns with lines. One column is the top 100 for a specific week, and each line represents one specific track. If the results of our analysis demonstrated 60%+ probability of the AI use, we highlighted this song with a specific colour. Obviously, it doesn’t always imply that the track was completely created by AI. It shows the probability of the fact that each specific song was created using AI-instruments or sounds as if it were created using them.
If you point at a song line, you will get detailed information and a link to listen to it, so that you could decide for yourself whether it was created by AI or humans.
AI-generated music on Ukrainian YouTube
Here you see the most popular songs on Ukrainian YouTube since the beginning of 2025. One column is the 100 most popular songs in a week, and one line is one song. We used colours to highlight the ones, the creation of which, most probably, involved the use of AI (with the probability of 60% or more, according to the analysis by AI Song Checker). Move over it for more detailed information, including the estimate of the probability of AI use from the NUAM experts. If you are interested in specific artists or songs, use the search. Click on a line twice (!) and listen to it yourself to see whether it was written by AI or a human!
AI-generated music on Ukrainian YouTube
Here you see the most popular songs on Ukrainian YouTube since the beginning of 2025. One column is the 100 most popular songs in a week, and one line is one song. We used colours to highlight the ones, the creation of which, most probably, involved the use of AI (with the probability of 60% or more, according to the analysis by AI Song Checker). Move over it for more detailed information, including the estimate of the probability of AI use from the NUAM experts. If you are interested in specific artists or songs, use the search. Click on a line twice (!) and listen to it yourself to see whether it was written by AI or a human!
Can one discern the AI-created music on one’s own?
Kateryna Yevgrafova has been teaching singing, composition, and music theory for 15 years, and she is convinced that professionals can discern AI-generated music by ear.
“However, frankly speaking, artificial intelligence is developing, and it becomes ever more difficult to catch these details. I have formed several criteria for myself, including the sound of the “artist’s” voice. In some places, it sounds like a robotic one. That’s how sometimes we hear the voice of an actual person who can’t sing very well and whose voice was “tuned in”. A computerized effect can be heard in transitions among notes,” Kateryna Yevgrafova explains.
In her video blog, she reviewed several artificially created tracks that entered the top charts, and among them, she analyzed the “Zero” composition by Amaya Roma.
“I heard very low-quality primitive music of no artistic value […] It reminded me of weddings from the 1990-s when everyone is already drunk and it doesn’t matter which music is used for dancing. I still don’t understand why it is in YouTube tops, because it has a very simple sequence a succession of chords in one tone, the melody contains repeats. I am not surprised that this music was artificially created, but if it were written by humans, I would have questions,” Kateryna Yevgrafova says.
The fact that the AI-generated track can be distinguished by ear is also confirmed by Maksym Komliev, the A&R manager of pomitni a music label, established on the basis of MOZGI Entertainment. It promotes Ukrainian artists, including Otoy, Dorofeeva, Nazva, and Anna Trincher. “This is a matter of experience, attention, and a trained ear. But, as a rule, AI-generated tracks are identified by repeated motives, a linear structure of sound, and an obviously unnatural voice. And in general, they sound too ideal, too polished,” Maksym Komliev explains for NGL.media.
In Maksym’s opinion, now artists with original songs, constantly competing for the attention of listeners, have to compete with the creators of artificial content as well. Yet, here is where the opinions of show business representatives differ.
For instance, Dmytro Zeziulin, the frontman of the rock group Latexfauna, is convinced that the boom in AI-generated music poses risks only for those artists who seek superficial and banal musical solutions.
“If your art is governed by the principle to do something innovative and creative, then you have to do it yourself. Yes, I have seen that there is a lot of AI on YouTube. For God’s sake, let it be there. Nobody will love Latexfauna less because of it,” Zeziulin says confidently.
How much do they earn from artificial music
It is not known who is actually behind the creation of the track “The Mist Dances Across the Earth” or Transportna group. NGL.media sent inquiries to the official emails of these artists, as indicated on YouTube, but received no response. The producers and music critics we have talked to have no idea who that is either. Yet, it is definitely clear that the affection of listeners brings the content owner quite a nice profit.
In the words of Andrii Komar, the CEO of Black Beats label, on average, one million views/streams on YouTube brings EUR 1,000-1,500, and Apple Music pays EUR 2,000-3,000 for the same amount.
“These are the prices we get as the Black Beats label; they can certainly differ from other channel owners, but on average, our figures are like that,” Andrii Komar explains.
To start earning with YouTube, it is sufficient to register on the platform called YouTube Partner Program. Yet, the site doesn’t have a detailed specification for profits, since they will depend on many factors (topic of the video, footage, geography and age of viewers, number of advertisement blocks), so you will get the final specification via individual email, once you join the programme.
It is relevant that the determination of the royalty amount covers both views on YouTube and streams on YouTube Music.
How is AI-music regulated on YouTube
Last year, YouTube issued an obligation for its users to label the AI content. This July, the platform promised to block the monetization of AI-generated video and audio. However, it remained unclear how exactly it would be determined. NGL.media asked Google representatives in Ukraine to comment on this issue, but we haven’t received any explanation as of the moment of this publication.
According to the specialists we have spoken with, the current situation regarding earning money on YouTube using AI-generated content appears to be unregulated. The rules of the platform prohibit receiving income for AI content, but until the account owner is caught doing it, they can still earn from it.
According to Ukrainian legislation, only a human being can be a creator and owner of copyright. So, it is important to determine whether this was musical content, generated by AI without the immediate involvement of humans, or if this content was created by a person using AI. For instance, a person could write an original melody and use AI for its instrumentation or just as a source of inspiration.
“Of course, the line between the generated objects and created art objects may raise disputes, but these disputes are already present in terms of authorship of similar compositions, created by humans without any AI,” Olena Andrienko, a member of the committee for the development of the artificial intelligence sphere at the Ministry of Digital Transformation of Ukraine, explained to NGL.media.
If this is an AI-generated composition, a person cannot specify themselves as its author, but can claim property rights for the composition, including the right to permit its use or to grant this permission to others. In Olena Andrienko’s words, regardless of whether AI is used for the entire generation of the content or only as an additional instrument in the creative process, it is necessary to read the licence conditions of the programme, in which it is created, very carefully.
Who may be harmed by AI-music?
Oleksii Yefymenko, the founder of the NUAM platform, believes that the prevalence of AI will cause the most damage to listeners who become accustomed to poor content.
“We have been living in a world where AI is everywhere, including music, for about a year now. It will not create any competition for artists. I believe that listeners are the ones to face the main risks. Unfortunately, our listener likes low-quality music, and now there will be even more of it,” Oleksii Yefymenko believes.
As for Oleksii Bondarenko, a journalist and a musician, he assumes that the interest to artificial music will soon vanish at all.
“Do you recall the time when everyone was afraid of a deepfake with Zelenskyi, that we will not see through where the real president is and where he isn’t? But it didn’t work. There was a fear that with the advent of Photoshop, we would no longer know which photos were real and which were not, but we dealt with it effectively. I assume it will be the same here, soon the outburst of interest in AI-music will pass away,” Bondarenko believes.
“I am sure that it could be regulated if it were profitable to someone. All these tracks should have their digital footprints, and in a few weeks, one could develop software to identify AI-generated music. Then YouTube could label these songs or lower their rating. But nobody has done that and might never do, because this music became popular. It is simply not profitable for YouTube to block this content because listeners will go to another platform – Apple Music or Spotify,” Oleksii Yefymenko concludes.
Mariana Verbovska (text), Nazar Tuziak (data analysis and visualization), Oleh Onysko (editor), Nelya Plakhota (translation), Viktoria Demchuk (cover)
