Geoff Taylor, chief executive of the BPI and Brit Awards, wrote this foreword for Music Ally and the BPI’s recent Music’s Smart Future report, addressing the potential of machine-learning, chatbots and artificial intelligence (AI) composition technologies.
“For most of us, the notion of artificial intelligence may conjure up visions of replicants from Ridley Scott’s dazzling sci-fi thriller Blade Runner, or the more recent humanoid offerings in Ex-Machina or C4’s Humans.
Yet AI is no longer the province of science fiction. In many areas of life, computers are now taking decisions or undertaking tasks (such as medical diagnosis or driving cars) that we previously assumed required human judgement and intelligence.
The technology has the potential to revolutionise the way we live and to shape the nature of our society. It also raises profound questions about the very essence of consciousness, our identity as humans, morality, and, yes, the Meaning of Life itself.
Thankfully those broader questions are beyond the scope of this report, but we feel it is time for us to reflect on the impact that AI technology may have on the music industry – and on music itself.
Recorded music is a sector whose DNA is closely entwined with technology, having been born from the invention more than a hundred years ago of the phonograph. Music companies have learned that while new technology throws up regular challenges to patterns of doing business, embracing new inventions to offer fans new and better ways to enjoy music is the long-term path to success.
That’s why the BPI has been running regular Insight Sessions on emerging technologies that will affect the music ecosystem, from blockchain systems, to virtual and augmented reality, to artificial intelligence.
As a creative technology business, change is our stasis. Recorded music is in the middle of its latest transition right now: for many consumers downloads replaced the CD, but streaming is now moving towards becoming the dominant paradigm in the music industry, last year accounting for nearly a quarter of UK music consumption – and that excludes all the video streams on sites like YouTube.
Streaming is clearly providing fertile ground for the application of artificial intelligence to music. The first area in which this is obvious is the growing importance of playlists and music discovery or recommendation features. We are entering an era of hyper-personalisation, in which consumers expect services tailored to their own tastes and preferences.
Early debate among music commentators pitched the merits of human curation against the power of algorithms, but this binary view missed the point that, here as elsewhere, humans can use the power of big data to better understand human behaviours and provide a more individualised experience.
Spotify, Apple Music and Deezer all use AI to analyse users’ behaviour and to analyse the relationship between songs so as to mix in new relevant tracks among a user’s favourites. And a new strain of AI technology is rapidly emerging that can take much greater account of contextual data: Google Play now mixes signals like location, activity, and the weather with machine learning about a user’s music listening, to try to provide the right song at the right time.
The human-computer interface is also developing quickly, with AI powering a new generation of “smart assistants”, such as Apple’s Siri and Amazon’s Alexa, which understand human commands and make it easier and more intuitive for consumers to access music online. These assistants are likely to evolve into our de facto musical concierges around the home and in the car (which of course they will be driving too).
Labels and artists are also embracing AI to engage fanbases. Bastille and their label Virgin EMI used a Facebook’s Messenger chatbot in a brilliant campaign to promote the release of their new album Wild World, while Robbie Williams and Olly Murs have created their own chatbots to engage fans. And label A&R and marketing teams are using “big data” analytics to predict consumer reactions and shape campaigns.
At the BPI we believe that gut instinct, passion for the music and human experience remain fundamental qualities in A&R and marketing, but as a sector we should not ignore new tools that allow us to reach fans in innovative new ways or which enable us to better understand complexity.
Finally, what does artificial intelligence mean for the most important aspect of all, the creation of music and for musicians? Academic researchers have been using AI for some years to create generative music but now major tech companies, such as Google and Baidu, have created programmes to create music using algorithms. We are seeing the beginnings of AI-composed music being used commercially, for example as background music in games and apps, such as by the UK start-up JukeDeck.
Some may fear that this will mean the sheet music is on the wall for human composers, and that we will all be consigned to a dystopian future surrounded by soulless muzak. But we believe that while there may, in the short to medium term at least, be some impact in low value background music applications, the insurgent technology is likely to be embraced by musicians and used creatively.
There are already early signs of this, with Brian Eno using AI to create video for his song The Ship, and 65daysofstatic utilising AI to produce textures for games for original fragments they created. There is a long way to go, however, before AI can create music as meaningful for humans as that arising from the random serendipity of the meeting of Lennon and McCartney, or the revolutionary spark of genius in classical composition that Mozart brought to the world.
For the longer term, AI raises fascinating questions. Can music created by AI inspire human emotions, engage the soul as human music does? If AI machines pass the “technological singularity” and become independently conscious, how will their compositions differ because of their different experience of what life is? At that point, will they be accorded rights, including the rights in their own creations? Could AI machines even become music fans and consumers?
Ray Kurzweil at Google predicts that this singularity will occur by 2045. Even if he is right, we believe that fears of AI taking over all music creation and fan engagement and rendering human input obsolete – music’s equivalent of Skynet in The Terminator, if you like – are pure science fantasy.
Music is a fundamental expression of humanity, and humans will always want to use it to communicate their ideas and to connect to one another. AI is the latest in a stream of new technologies that the music business can use to enhance the connection between artist and fans. It’s time to embrace music’s smart future.”
Download the full Music’s Smart Future report as a PDF, or read our coverage of the BPI insight session that accompanied its release.