The digital revolution has brought pain and promise to the music industry. Now, artificial intelligence (AI) looms on the technological horizon as the next great disruptor.
Machines that write songs? Software that scores music for film and corporate videos? They’re not some distant sci-fi novelty, they’re already here, assisting human composers – but also threatening to replace some of the work previously done exclusively by humans.
Last year, researchers at Sony Computer Science Laboratories released a song in the style of the Beatles, called “Daddy’s Car,” created by the AI application Flow Machines (a research project funded by the European Research Council, and co-ordinated by Sony CSL Paris).
French startup Aiva (Artificial Intelligence Virtual Artist) is an AI composing app rooted in classical music and aimed at the synch market.
According to Aiva’s creators, “We have taught a deep neural network to understand the art of music composition by reading through a large database of classical partitions written by the most famous composers (Bach, Beethoven, Mozart, etc.). Aiva is capable of capturing concepts of music theory just by doing this acquisition of existing musical works.”
Aiva’s first album, Genesis, is available on Soundcloud. And this year Aiva became the first composing app to do a deal with a performing rights organization; all music composed by Aiva is automatically registered with SACEM.
And while smaller AI music startups are sprouting like dandelions in spring – witness some of the ingenious applications on show at the Techstars Music Accelerator in Los Angeles – Google, Microsoft, IBM, Apple, and Amazon, are all investing heavily to stay at the forefront of the AI revolution.
Last year, Google launched Magenta, a research project aimed at pushing the limits of what AI can do in the arts. This is done using artificial neural networks – computer systems inspired by the neural networks of the human brain.
“The deep-learning revolution comes courtesy of the computer-game industry,” writes Larry Hardesty, in MIT News. “The complex imagery and rapid pace of today’s video games require hardware that can keep up, and the result has been the graphics processing unit (GPU), which packs thousands of relatively simple processing cores on a single chip. It didn’t take long for researchers to realize that the architecture of a GPU is remarkably like that of a neural net.”
Arne Eigenfeldt sees AI as a creativity booster, rather than a replacement for human creation.
IBM is exploring a range of music-related AI applications, including Watson Beat. Explains IBM researcher Kelly Shi, “Watson Beat composes music by ‘listening’ to at least 20 seconds of music, and then creates new tracks of melodies, ambient sounds, and beats based on what it learned from the original sample – whether the user wrote it, or is using other samples and songs.”
SOCAN is working with IBM Watson and Canada’s leading AI labs to utilize the technology on behalf of songwriters and music publishers. “SOCAN is very committed to artificial intelligence,” says SOCAN Chief Operating Officer Jeff King. “We want to be a global leader in this area.”
King says AI is particularly useful in music identification. “Watson can look at 700,000 web pages per second,” he explains. “We’ve applied the power of Watson to YouTube and user-generated content. Initially we focused on lyrics: Watson would learn a song then go looking for matches. We had good success. Then we went looking at the music, melodic patterns and so forth, and Watson did reasonably well. But when you combine the two functions, examining [both] lyrics and melody, then the probabilities were much higher. So we’re seeing AI as a really interesting opportunity to license and monetize cover versions in a smart, non-manual way. Using AI like this could really transform the industry.”
King adds that AI has other positive applications for SOCAN. “We can use the processing power and learned logic to identify when someone is starting to break out,” explains King. “For instance, in our Watson experiment, we discovered that when an artist is mentioned on social media outside their postal code, they’re likely on the verge of doing something. We’re using things like that to help our recruitment activities, to identify people who should be part of the SOCAN universe.”
Vancouver-based composer and Simon Fraser University professor Arne Eigenfeldt, is dedicated to exploring metacreation (imbuing computers with creative behavior); Eigenfeldt is an expert on the subject of musebots, which are virtual musical agents that make music together.
Says Eigenfeldt, “Most of my music in the last 10 years has used AI in some capacity, and I’m definitely part of the worldwide computational creativity, as well as the musical metacreation, community. Both are concerned with automating the creative process through computation, i.e., ‘using AI to make art.’ Or ‘artificial creativity.’ Or ‘machine creativity.’”
Eigenfeldt sees AI as a creativity booster, rather than a replacement for human creation.
“Computers are tools for artists, and allow us to do things much more easily than before,” he said in an e-mail. “More powerful software on these computers will allow us to accomplish things in much shorter time, but also in new ways. Prior to my exploration of AI in music-making, I felt I was in a creative rut, relying upon the same ways of working that I had for years. Now, my software is a creative partner that allows me to think about musical creation in ways I never would have imagined.”
There’s no question that AI will have a profound impact on the landscape of music creation. But AI is also being used to discover and recommend new music, an important influence in a streaming world with millions of songs to choose from.
Earlier this year, Spotify acquired AI music startup Niland in a bid to improve music discovery and its music recommendation back-end. Leading music data identification company Gracenote has also invested in AI in an effort to better classify mood and emotion in songs.
And AI’s application may help to usher in a new era of data analysis resulting in improved royalty tracking and payment for all music rights owners. It could revolutionize the way we monitor billions of small transactions and data exchanges in the digital world.
As promising as these developments are, AI is merely the tip of the technological iceberg portending further upheaval and creative gains in the years ahead. Says Eigenfeldt, “Our notions of creativity may evolve because of these new tools, but if we evolve as well, so will our art.”