Roman Ryabtsev: “Musical Ragnarok, the extinction of a creative person”

Regular readers of my column have already noticed that making predictions in the entertainment industry (of which music is, of course, a part) is one of my favorite things to do. But few people know with what malice I (to myself, of course), wearing a T-shirt with the inscription “I told you so!”, Like to read in the news another confirmation of my prophecies used the pronoun “I” only six times!). So, now I’ve found myself a visionary in the field of AI, the notorious artificial intelligence, and its presence in the music business. Surprised? I am not at all.

Less than a year (only ten months) has passed since my article in ZD under the heading “Robot guitar competitions are going on”, where I predicted that artificial intelligence (hereinafter referred to as AI, as is now generally accepted) will soon take orders from living composers in the field of writing background music for metaverses, supermarkets, malls and other public spaces. Let me quote myself (as befits a person of universal modesty): “According to my forecasts, the first such AI composers will be available in the metaverses in about a year and a half. And in two or three, they will capture the majority of the background music market in virtual worlds.

I have to admit that I was wrong in my predictions. No, not in evaluating the possibilities of AI in the entertainment industry, but in timing! The latest news and events in the field of IT have outstripped all my wildest assumptions: the future has already suddenly come and continues to step on the heels of leather bags, which still proudly call themselves composers, musicians, singers, artists, poets …

Let me guess: surely you, in the last month or two, have had fun creating strange pictures using various neural networks? Or did you repost on social media a collection of funny art produced by AI, which people give all sorts of crazy tasks, for example, draw Star Wars or Harry Potter characters as babies? Six months ago, this was practically not observed on the Internet, and now the capabilities of neural networks have massively rushed forward and upward (because they have obviously already accumulated fairly large databases and have trained to fulfill human wishes). And now computer art is available to absolutely everyone, even an ordinary smartphone is enough to get absolutely amazing pictures at times at your request.

And the other day I came across a release of some group on the Web, I look at the cover … And I clearly understand that it was not drawn by a living artist, but clearly generated by a neural network! Well, it can be seen so far by some signs, especially if it’s done in a free application. But even now this cover looks better than most of the Russian artists. So, gentlemen, designers, soon you will have to look for other areas of application for your talents: computers do not require an advance payment even before creating sketches, do not go into a creative binge and do not throw tantrums “I am an artist, I see it that way!”

“Well, no, well, in the field of songwriting, but such that the soul unfolds and then curls up, no neural network can certainly cope! Well, how can she know about real Feelings, Soul and Creativity?..” – you will say. Well, I have very bad news for those who firmly believe in the indispensability of Homo Sapiens.

Recently, one of China’s largest streaming platforms announced that its Lingyin Engine AI, which sings with the voices of various artists (which already sounds fantastic), has sung over a thousand songs in the past three months. The most successful hit by this “soulless program”… received over one hundred million views, bringing the company almost $350,000! Lingyin Engine invites the authors of music and lyrics to order the performance of their songs by the voices of various stars (both living and not so). For example, Elvis Presley himself can sing your imperishable song! I wish I could hear White Roses performed by him… Or American Fight… Or something from Morgenstern, Cadillac, for example… I will modestly keep silent about Strange Dances… True, now the company has limited access to the instrumentation and releases songs only with voices Chinese singers (living and dead in Bose).

This AI can create an infinite number of songs in any language and in any musical style: a computer does not need to spend years learning music, learning to play musical instruments and vocals, it is enough to analyze the cultural heritage already accumulated by mankind, and … the music market can change and forever. For such streaming companies that spend millions of dollars annually on payments to music performers, the new trend for AI artists can become a real El Dorado: the neural network does not ask to eat, but only generates income.

You know, I was not too lazy, I searched and listened to the songs performed by this Chinese AI. Now, if I didn’t know that it’s not a living person who sings, but a vocal synthesizer … I wouldn’t have guessed in my life: everything sounds so flawless! If there were a Turing test for music, then this AI would definitely pass it.

And this is happening all over the world: in South Korea, songs that are (still) performed by live musicians with music and words by AI are popular. Capitol Records signed a contract with virtual rapper FN Meka (however, he also only writes lyrics and music, but a person performs), and the music algorithm of the Endel application generally signed a contract with Warner Music to release as many as twenty albums!

Yes, and we have enough similar projects in Russia: from imitation of Scriabin to … For example, one AI writes texts in the spirit of Yegor Letov, and this is released under the name “Neural Defense”. Personally, I didn’t notice the difference between the machine-made verses and Letov’s original songs at all (the sarcasm mode is on: well, I’m not a connoisseur of Civil Defense, sorry).

✭✭✭

There are many more examples of successful AI artists (I wrote and thought: how crazy it sounded a year ago!), But all sorts of legal problems associated with music and performance generated by neural networks have already risen to their full potential. For example, in the USA there is a very funny legal conflict: in American copyright laws, the word “human” never occurs! So AI has the same rights? Lawyers are scratching their heads. And in the same USA back in 1965, the Department of Copyright Protection recorded the received application for computer-written music. But nothing has changed in the legislation since then, and American lawyers do not know what to do with virtual authors now.

For example, if an AI analyzes the entire work of, say, Madonna, composes and records (and releases, as yet with the help of humans) an album that sounds indistinguishable from the real Louise Ciccone, would it be considered plagiarism? Can the real Maddie claim a share of his sales since HER voice is heard? Some lawyers believe that this will be plagiarism, others that composing songs under the influence of someone else’s creativity is a common thing in the world of live musicians, and it seems that no one is punished for this (in Russia, for sure, judging by the number of clones of “Korol i Jester” on our rock scene).

By the way, the concept of “person” is absent in Russian legislation! “Licensor”, “licensee”, “author”, “performer”… Nowhere is it written that they are made of flesh and blood (take my word for it, how many contracts I have signed in my life). So, sooner or later, our legislators will have to pay attention to the growing threat to living authors from neural networks.

And just the other day I saw an unrealistically touching Christmas video that was composed, drawn and voiced by neural networks. The story of a donkey who dreamed of becoming a deer in Santa’s team turned out to be so HUMAN that if I hadn’t read in the accompanying text first what AI did it, I wouldn’t have guessed it in my life. And what a soulful voice the announcer had, who read this tale … but it was also a computer. And this sweetest fairy tale sounded like a real threat to voice actors all over the world: in every country, films are dubbed in the local language … But it will still take a little time, and the viewer will be able to select the language spoken by the characters in the film directly from the TV remote control (as we now select subtitles): AI will voice it with ORIGINAL voices of actors in any language of the world. The need for dubbing will soon disappear completely, and some actors (who have seen this video) already understand this and are seriously thinking about looking for other earnings in the near future.

I sympathize with them, but … I would like to see the same “Harry Potter” or “Iron Man” with the real voices of the actors who played them, but speaking Russian, just imagine it! ..

To summarize: the successes of AI in the field of musical and other creativity and the wave of serious legal and purely human problems raised by them look like an impending Ragnarok. A sort of last battle between living people of art and an army of computers, which are better and better able to cope with the task of spiritual nourishment of the broad masses of the people. But in Scandinavian mythology, Ragnarok was not the end of the world: it should become the very battle in which the old will perish and the new will be born. How humanity will cope with this Neuro-Ragnarok, I will not venture to predict yet. Although I really like to make predictions, yeah …