AI will pick (and make) the songs you love in just 6 years

Music has a rich history of technological and digital augmentation. Think about Peter Frampton and Jimi Hendrix using “talk boxes” to make their guitars talk for them. Or how about the billions of people influenced by the computerized sounds of pop music from Michael Jackson, Prince, and Madonna. This is partly why the music industry is already accepting a revolution in artificial intelligence.

Music is an industry where the technologically unsavvy musicians can’t necessarily keep up. A musician without access to studios, equipment, and high-tech software may be able to rock a small live audience. But they cannot produce and master their sound to national syndication quality.

Compare this to other art disciplines. Visual artists can use oil and canvas and still be pictured in the Museum of Modern Art. Writers can type on 1950s typewriters and still influence the entire country with their prose. The equipment makes no difference.

This technological barrier to entry in music precipitates into a massive opportunity for artificial intelligence. It’s why we’re seeing the adoption of AI in the music business unparalleled by any other art form.

Take, for example, LANDR, which is an AI audio mastering tool that’s disrupting recording studios. Essentially, LANDR takes on the crucial role of sound engineer, helping artists master their sonics. With millions of analyzed sounds in its archive, LANDR can also help artists find the sound unique to their identity. Best of all, it’s accessible (all online) and affordable (just a few dollars).

Sound engineering isn’t the only role AI entering either. It’s also behind the instruments.

Amper is an artificial intelligence composer, performer, and producer that empowers artists to instantly create and customize original music for their content. As far as I know, they’ve got a major hit under their belt with an artist named Taryn Southern. And many more to come.

Similarly, AIVA is an AI composer making major progress in music production, however, with a different approach. Instead of catering to artists, they want to be the personalized musician for everybody (particularly gamers at first). They recognize that there are many moments and emotions in life that lack music. AIVA fills this by making great quality compositions personalized to the listener.

We’re going to see more programs like Amper and AIVA emerge that’ll give pop artists a catchy loop, rap artists a hard beat, and every listener a song just for them.

When Kanye West introduced the Hip-Hop world to Auto-Tune on his album 808s & Heartbreak, the world was appalled by the technology. Later, T-Pain would bite the style and create an entire career around the tool. And now nearly every hip-hop or pop artist uses the tool in some capacity on a regular basis. AI music production tools are very much in the same boat.

Will any of these AI tools single-handedly create the next Thriller? Not on their own. But, it lowers the barrier to entry and puts the possibility to create magic in the hands of more artists. For this reason, AI tools are going to drastically elevate the quality and variety of music that makes it to the mainstream.

By 2026, 55% of the Top 40 Billboard singles will be partially written or produced using AI tools.

Today, Spotify sees about 20,000 new uploaded songs a day. This equates to over 7 million songs a year. And that’s just one streaming platform. There’s also YouTube, SoundCloud, Apple Music, etc.

This number of music creation will only amplify over the next years with this widespread access to professional production tools.

How can we, as consumers, expect to weave through all of this material to make sure we still discover the great music? How can record labels expect to weave through all these artists and make sure they’re offering opportunities to the right ones?

Already a future thinker?
Then become a friend.

The Bad Side of Data

Take any historic record label and you’ll find at the heart of their success is a great A&R (artists & repertoire) Executive. They are the lifeblood of a label. They are responsible for spotting emerging talent and identifying where the industry is headed. In a way, they are futurists of the music industry.

When an A&R is on top of their game, like L.A. Reid when he signed the 14-year-old Usher or John Hammond when he discovered the 18-year-old Aretha Franklin, they influence entire genres of music.

When an A&R is off their game, like RSO Records passing up on U2 or the numerous labels that didn’t care for Jay Z, they might be responsible for their label’s demise.

Because their job deals with recognizing new sounds and not just riding the quick trends, A&Rs always tended to have a great ear for music and the cultural zeitgeist. Then, about fifteen years ago, with the advent of digital music platforms, record labels replaced their musically-inclined A&Rs with data-minded businessmen. The folks in charge of supplying the record labels with fresh talent began thinking far more about metrics than actual sonic influence. It became a game of marketability – where image and branding outweighed talent.

This is largely where the industry is today. Labels give their biggest budgets and opportunities (radio spins, marketing, etc) to artists whose images they can control and get clear data on. I don’t believe this is for a greater cultural conspiracy. Rather, they want to make sure they make money. Therefore, taking risks isn’t how it’s done.

It’s all about to change, though, with the emergence of the AI-powered A&R.

The Good Side of Data

Because music has gone digital and streaming platforms collect so much data, the opportunity for intelligent systems to analyze this data with greater, thoughtful scrutiny is here.

The AI-powered A&R is an intelligent system that analyzes music streaming data, fan engagement on social media, and other data to discover the next big talents in music along with gaining strategic insight into the artists they’ve already signed.

Some of the more popular AI-powered A&Rs are Asaii and Instrumental. There’s also Sodatone which was bought by Warner Music Group.

How does the use of AI-powered A&Rs differ from the last couple of decades of data dependence?

The difference here is that they’re taking into account more streams of data and the models for analysis are beginning to reach their true potential. This shifts the data discussion to emphasize quality over quantity.

For instance, let’s say they’re analyzing two new artists: Alphonse and Bambino. Alphonse has 500 fans but consistently engages each for an average of 35 minutes a day (10 songs). Bambino has 6,000 fans but on average engages them for only 12 minutes a day (3 songs).

Even though Bambino gets more than five times the stream time as Alphonse, in theory, an AI-powered A&R would choose Alphonse over Bambino.

This is because the data shows far more fan engagement for Alphonse, who’s defined his audience persona. He knows the people that love his music and now it’s all about finding the others just like them. Bambino, although with a bigger audience, is a little more discombobulated as far as finding his ideal audience.

Ideally, just a few years from now, an AI-powered A&R is to be able to find the next hit song when there are only 65 listeners. Contrary to the A&R today who might not discover the song until it’s at 1 million hits.

I know it’s a cliche’ to say that the AI will help humans, not replace them. However, this is truly the case for an AI-powered A&R.

If we refer back to the number of new music generated every year (7 million songs on Spotify alone), it’s clear that finding quality talent depends on breaking through the noise. This is where the AI-powered A&R will excel – analyzing thousands of new songs and artists a day to narrow the field of emerging talent. This should give record labels the chance of bringing back some of the human A&Rs that have a great ear for music and what the next sounds to influence the culture will be.

Who knows, with AI composers like Amper entering the field of music, one of these AI-powered A&Rs might one day sign an AI composer to their record label. Wouldn’t that be ironic?

join the discussion