Is it harmonious not to know that we are competing with AI?

Reflection: Is it harmonious not to know that we are competing with AI?

A brilliant student and a better professor on Big Data and Business Analytics at IE University shared with me the way Spotify classifies music. If you want to position any single song on the music platform you have to be familiar with several taxonomies and audio features such as energy, mood, danceability, genre, but also rhythm, tempo, valence, loudness, liveness and so on.

Anytime a composer uploads a song on to Spotify, the music platform sets a score in all those taxonomies and audio features. Apparently each song has a single score resulting from the weighted average of different time slots in each and every audio feature [1]. That’s the way I learnt from Pedro Figueiras how Spotify translates music into data.

If music can be translated into data, could data (natural language or code) be translated into music? Yes it could and it can. That is the case of MusicLM, Jukebox and so many other AGI solutions that are on the horizon in the near future.

Back in 2019 Huawei launched an experiment [2] to complete Schubert’s unfinished symphony No. 8 and the completed symphony was interpreted at Cadogan Hall in London in February 2019. Astonishing, isn’t it?

I have talked to classical music instrumentalists and they think the composition has no soul although it resembles something similar to a Schubert’s creation. Whether we feel comfortable with this music drift ir not, we have to admit that right now it’s a reality.

Next step in this journey is that now anyone could even compose music without speaking the language of music. You no longer need to know solfeggio to compose music or at least something that sounds like music to the ears of the uninitiated (the majority of music consumers).

Here is my reflection. Last summer Spotify acquired Sonantic for €93 million [3]. An AI startup responsible for creating real human voices. Isn't the voice the best instrument? If Google and OpenAI has already released text-to-music solutions, If Spotify translates music compositions into data sets, if Spotify remunerates with royalties music bands and composers based on the number of plays, if Spotify is a for-profit corporation, how long do you think Spotify has been releasing its own music groups and synthetic music compositions made with AI to optimize its cost structure?

I don’t know the answer to this question and I know for sure that there should be a limit for undercover AI competing with humans in any cognitive domain. Jack Kerouac said and I quote "the only truth is music" and now I wonder if what we are creating counts as truth. Let humans be the judge.

"The only truth is music" –Jack Kerouac

By the way, this same reflection applies to the creation of content (posts and videos) on social networks, articles in newspapers, books on Amazon, image and design in advertising, code on GitHub, and the like.. We are on the threshold of a new synthetic era in which we are gripped by AI-generated content thanks to platform-owned AGI solutions to save royalties and costs. Is that sustainable?

Shouldn’t we be informed upfront when the content we are consuming is no longer UGC (user-generated content) but AIGC (AI-generated content)?

What’s your thoughts on this issue? Happy to hear your comments on the thread.

[This content is written by a 100% human being]