Can machines make music? Google tests creative boundaries of AI.
Google's Project Magenta uses algorithms to try and develop music, but critics question whether a machine could ever really create art.
The latest venture from Google's innovation team has many in the creative community pondering an old question: What is art?
Magenta, the internet giant's latest artificial intelligence project, is a stab at programming an algorithm to create music.
Magenta has already produced a single "song" of sorts, a four-note melody composed by Magenta and accompanied by a human-produced drum beat.
"Machine learning has already been used extensively to understand content, as in speech recognition or translation," wrote Magenta researcher Douglas Eck in a blog post. "With Magenta, we want to explore the other side – developing algorithms that can learn how to generate art and music, potentially creating compelling and artistic content on their own."
Will it work? If art is a creation of beauty and a means of sharing a message, then Magenta has struck out so far, as reviews have been scathing.
"It’s not totally awful," wrote Alexander McNamara for the BBC's Science Focus. "Actually, we’re not sure even Adele blurting her lungs over the top of it could help this number in the charts, but as Google [points] out this is just the beginning."
Others questioned whether anyone would actually want to listen to this music after the novelty wore off.
"Google may have annihilated its human competition in the ancient game of Go, but human artists don’t have to worry – yet – about losing their livelihoods," wrote Mark Hachman for PC World.
The criticism is partly just relief. If machines had already advanced to the point of communicating emotive experience through song, then how far away are robots who write technology articles? Machines have already overtaken many of the labor positions that once fueled the economies of entire cities, so a machine attempting the thoroughly human creation of art is concerning to many.
"If machine learning can create hit songs or beloved stories at the push of a button – which would amount to passing the Turing test – the end result will no longer be art," wrote Thomas Claburn for Information Week. "It will be commoditized content, a new flavor of the drivel dispensed by content farms to capture ad revenue."
The Magenta team insists the research is a matter of better understanding learning. The project was inspired by Google DeepDream, where programmers "taught" a neural network to recreate surrealist images by showing it a series of photographs.
"There's a couple of things that got me wanting to form Magenta, and one of them was seeing the completely, frankly, astonishing improvements in the state of the art [of creative deep learning]. And I wanted to demystify this a little bit," Mr. Eck said during a panel at the music and tech festival Moogfest, according to Popular Science.
Magenta is being open-sourced, with the hope of building a research community of programmers, artists, and even human musicians.
"The question Magenta asks is, “Can machines make music and art? If so, how? If not, why not?” Eck wrote.