Would Mozart jump out of his grave if he knew what AI is doing to the music industry?

Reaching out to your fans today takes more than just good music, showmanship and live concerts. With the proliferation of social media apps and the many different kinds of streaming services from Apple Music to Tidal, Artificial Intelligence (AI) is playing an increasingly important role in letting ordinary fans discover new artists around the world across different genres, be it jazz, reggae, EDM or R&B.

By identifying certain patterns in the tempo and arrangement of the music we listen to on a regular basis, AI is able to predict what we would love to hear. Predictions equal certainties, which is why AI is so important in today’s world. Big data collected through social media and from our browsing histories on the internet helps music tech giants, like Spotify and Apple Music, to capitalise on our preferences and behaviour. Through these data and predictions, the ‘right kind of music’ could be introduced to us. Given the staggering volume of music streamed by these giants – Spotify alone sees more than 20,000 songs uploaded every day – AI drastically increases the efficiency and effectiveness of delivering the right kind of music to listeners based on what they have listened to in the past.

So is it a good thing to have AI create playlists based on our past data?

I think it is ‘okay’ but only to a certain extent.

While AI helps music streaming services to more efficiently deliver what we liked in the past, a songwriter like me would much prefer broadening my music horizon by listening to a wide range of artists. I love the music of Bruno Mars and Shawn Mendes, but I also like to listen to indie artists from time to time to get a fresh perspective or sound. In short, I want to listen to music of new kinds that has never been on my playlist.

What would Mozart and Stravinsky think of this AI revolution in music? If I were to make a bold assumption, I think both Mozart and Stravinsky would probably have embraced it but not without certain reservations. Why? Because solely composing music based on past successful patterns would create dull and boring outcomes.

In recent years, AI advocates have argued that computers can generate great popular music by analyzing data from different compositions. Algorithms can put together a winning formula by systematically “cutting” and “pasting” characteristics and patterns, thus creating music that is enjoyable or that mimics a certain genre. Yes, computers can do a lot, but there is one thing computers can NEVER do: create music based on real human feelings. Composers and songwriters delve into their real life experiences, emotions, and intimate feelings to tell all-too-human stories through music. Computers may mimic “notes” and “rhythms” but they do not and cannot fall in love.