Mubert, a company run by a software engineer named Alex Mubert, has announced that its artificial intelligence (AI) application has created 100 million songs. “That’s roughly equivalent to the entire catalog of music available on Spotify,” Ted Gioia writes on his online newsletter The Honest Broker. “The company notes that this adds up to 4.8 million hours of creativity.”
http://www.honest-broker.com/p/the-number-of-songs-in-the-world
(via http://www.artsjournal.com)
AI, which has been a thing ever since computers started winning at chess, is now The Thing in technology. Depending on who’s assessing its impact, it will either save or destroy human endeavor.
Proponents say it can kick scientific and medical research into hyperdrive, getting results in seconds rather than years. Critics worry that AI-enhanced robotics could upend a range of professions, from oncology to long-haul trucking, depopulating the workforce, de-humanizing the acquisition and use of knowledge – making most of us, as the British say, “surplus to requirements.”
Fields requiring human sensibilities, notably the arts, might have been considered immune to AI. A couple of years ago, software took assorted sketches by Beethoven and assembled a Tenth Symphony. Reviewing this creation for the online classical magazine Van, the Beethoven scholar Jan Swafford found that the result was “something that sounds unquestionably like a piece of music, only a gangly and forgettable one. . . .
“Artificial intelligence can mimic art, but it can’t be expressive at it because, other than the definition of the word, it doesn’t know what expressive is,” Swafford wrote. “It also doesn’t know what excitement is, because there’s a reason people call excitement ‘pulse-pounding,’ and computers don’t have pulses.”
The Intelligence of Bodies
While AI may be incapable (so far) of credibly composing a symphony, it may be usable for incidental and background music, jingles, ring tones and the like. AI’s prospects for composing pop songs are more murky.
Commercial popular music already is significantly automated in instrumentation (synthesizers, drum machines, etc.), and the compositional group-think that makes songwriters and performers replicate trendy sounds and subjects is as old as music itself. Everybody copies everybody else, and always has.
I haven’t heard Mubert’s songs. Maybe they will be indistinguishable from what we already hear on light-rock or smooth-jazz channels. Songs reflecting human feelings and life experiences, interpreted by humans with individual personalities and distinctive musical techniques – i.e., songs worth listening to – will be very hard to automate, even by the most sophisticated software.
Letter V 