Rise of the robot music industry
Rise of the robot music industry
AI is transforming music streaming, talent spotting,
promotion and even composition
December 2, 2016 by: Nic Fildes, Telecoms Correspondent
Robotic is not an adjective that many musicians would
want applied to their songs but the industry has been fast to embrace data
analytics and artificial intelligence to help tailor its services to the
increasingly fickle listener.
Algorithms are seeping into the music business to help
with talent spotting, promotion and even composition in an industry that has
been historically resistant to change and was one of the first to feel the
effects of “disruption” through piracy and music sharing.
Streaming services have already ushered in an era of
“hyper personalisation” for music lovers. Spotify’s Discover Weekly playlist,
launched in July 2015, had racked up 40m listeners around the world and 5bn
track streams by May this year, according to a report from the BPI prepared by
Music Ally. These playlists monitor what a person is listening to, and
cross-references that data with other users with similar tastes to recommend
new songs and artists.
Apple Music has opted to use human curators such as Zane
Lowe, the radio DJ, for its playlists, but Spotify has doubled down on its
robotic recommenders with new services such as Release Radar and the Daily Mix
to tempt its subscribers down different paths.
Yet discovery is only the equivalent of a debut album for
streaming services, and can be a blunt tool. Users of Spotify Discover complain
that it is hit and miss — often suggesting the same artists and songs
repeatedly, and failing to adapt to the often random whims of the listener.
The industry is now hoping that the use of artificial
intelligence will bring better analytics, and even predictive technology.
A listener’s location, mood and even the weather
conditions are now being built into some recommendation engines. Google Play
is, for example, working on such adaptive functions.
“A bot will be able to recognise guilty
pleasures . . . see that I’ve been to the pub and serve me a Little Mix record
when I’m on the way home,” says Luke Ferrar, head of digital at Polydor,
pointing to the use of algorithms to understand how people listen to music.
When combined with the sort of intelligence provided by a
smartphone — location, time, activity and movement — it means that music
services can find the right track for the right moment. In effect, AI can
determine whether a person is bored in an airport, studying in a library or
sunning themselves on a beach, to tailor a playlist.
AI has already started to be used to improve streaming
services. Quantone, a London-based music AI start-up, is using the IBM Watson
engine to further improve recommendations by crunching music reviews, blogs and
Twitter comments into how music is analysed.
Evan Stein, chief executive of Quantone, said AI allows
for a more precise data set than “you like Iron Maiden, you’ll probably like
Metallica” to one where someone who appears to like a certain bass player can
be pointed to other records featuring the same musician.
The rise of smart assistants such as Apple’s Siri and
Amazon’s Alexa in the home also points to a future where AI acts as a “musical
concierge” in the living room or car according to Geoff Taylor, chief executive
of the BPI.
AI’s role in the music industry is also expanding into
the business. Record labels have started to use “chat bots” — computer programs
that interact with consumers — to promote new albums and tours. The pop singers
Robbie Williams and Olly Murs have launched bots to answer questions from fans
and push them to buy more from online stores. Bastille, the British band,
created a bot that masqueraded as an evil company called WW Comms that sent
fans Gifs and video clips.
There is a lot of
hyperbole about robots taking over but its more about getting a better hammer
to hit more nails
Evan Stein, chief
executive of Quantone
There is also the opportunity to use AI to find new
artists. Instrumental, a British label, scrapes YouTube for people uploading
their songs and then sifts through data on thousands of unknown artists to
define which have started to attract attention. The label, which is backed by
Warner Music, has signed three of the most promising to development deals.
Some remain unconvinced that old-fashioned talent
spotting is set to be replaced, however. Simon Wheeler, head of digital at
independent label Beggars Group, told the Midem conference in June: “We have a
role of finding things that people don’t know they’re going to like . . . and
data are not very good at doing that stuff.”
The biggest question is whether the robots will start
making the music too. Google’s Deepmind has been used to create a piece of
classical piano music, while the technology company’s Magenta research project
is using machine learning to create “compelling art and music”. That leads to
the question of whether sophisticated machines will end up creating music for
their own enjoyment, according to the BPI. In other words, will androids dream
of electric guitars?
British start-up Jukedeck, which operates out of TechHub,
has already used AI to created half a million pieces of original music aimed at
companies and video creators looking to create fresh pieces rather than paying
royalties. This is hitting the stock audio industry and has the potential to
reduce royalties if retailers, for example, use Jukedeck to create muzak rather
than playing hits in store.
Mr Taylor said: “Some may fear 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 Ed Rex, co-founder of Jukedeck, does not think AI
will kill off the human composer, but instead expects more musicians to use
algorithms to improve their own work.
Mr Stein also remains unconvinced. “There is a lot of
hyperbole about robots taking over but its more about getting a better hammer
to hit more nails. A terrible composer will still make terrible music, just at
a faster speed.”
Copyright The Financial Times Limited 2016. All rights
reserved.
Comments
Post a Comment