Artificial intelligence is being used to create entire albums, and it’s a bit scary

Artificial intelligence is being used to create entire albums, and it’s a bit scary

Photo: Dhe Haivan
Words by Annie-Mei Forster

The future is here, whether you like or not.

Imagine listening to a whole album that’s been composed and produced entirely by artificial intelligence. Turns out it’s already been done. Last year US singer-songwriter Taryn Southern released her LP, I AM AI, and it was entirely composed using AI technology, not that you would’ve realised. Although it’s not a particularly groundbreaking album, it poses the question: where is the future of music heading?

Writing music using machines certainly isn’t new; in the 1950s, composer Lejaren Hiller used a computer to produce the ‘Illiac’ Suite for a string quartet, for example. But electronic music is only growing in popularity, so AI is just another tool that musicians can use to produce tracks. In fact, you’ve probably heard AI-produced tracks without even realising.

A number of tech startups are increasingly using AI to create music. Music tech company Jukedeck produces AI jingles — everything from classical to pop — and sells them for use in videos, games or commercials. The cost of an AI track is a fraction of the price of paying for a real musician.

The good news is that although AI can spit out music quicker than a human, it takes a lot of time and data for it to produce music that a person would actually want to listen to. Humans have the advantage of being able to improvise and put their own unique flavour, something a machine simply can’t do – for now.

A commonly used program among budding musicians is Amper, an AI tool used to create tracks even if you have little experience. With its easy to use interface, you can create music based on genre, duration and mood. Best of all, you can use it for podcasts or audio assignments given there’s no copyright.

Musicians know how costly it can be to hire recording studios and lugging heavy equipment around certainly isn’t fun. Using AI could make producing music more affordable, especially for artists and bands starting out. In turn, this would break down some of the financial barriers currently preventing access to the music industry

The laws surrounding the use of AI are still catching up and the issue of copyright is quite murky. If a musician produces music using AI, who do they credit? If an AI program creates a track using samples from another artist, can they then sue? Can you even sue a robot? This is a grey area and it’ll remain that way until governments put laws in place.

So how does AI actually create a piece of music? In one scenario, a person could start by inputting certain parameters like beats per minute, instrumentation, genre, key, etc. The platform then spits out a song and you, the musician, can make adjustments. An example of this is Google’s AI Duet, software that will jam with you.

While still in its infancy, AI is progressing so quickly that it’s taken experts by surprise. The advances AI has made in the last few years alone are both amazing and terrifying. Taryn Southern once said if you’d told her a few years ago she’d produce an entire album using AI, she wouldn’t have believed you. As the saying goes: if you can’t beat ‘em, join ‘em.