björk has collaborated with microsoft to create AI-generated music based on on the changing weather patterns and position of the sun. called ‘kórsafn,’ which means ‘choir archive’ in icelandic, the composition will be played continuously in the lobby of sister city, a hotel in new york’s lower east side that opened in the spring of 2019.
kórsafn uses sounds from björk’s musical archives, which she has compiled over the last 17 years, to create new arrangements. it includes that of the hamrahlid choir from iceland, which björk used to be a member of and featured in her experimental theatre show, cornucopia.
images courtesy of microsoft
using microsoft’s AI, the music will adapt to sunrises, sunsets, and changes in barometric pressure, using a live camera feed from the sister city roof. the result is an endless string of new variations that creates an all-encompassing mood for guests. the ongoing computer vision project will train microsoft’s ai to better recognize dense and fluffy clouds, snow, rain, clear sky and birds in different lighting and seasons.
‘an architectural structure downtown manhattan offered me the hand in an AI tango and i accepted the call, I am alert with curiosity waiting the results’, says björk in a statement. ‘I offered them my choir archives, written over 17 years that will float through the pinball of artificial intelligence by the grid of bird migrations, clouds, aeroplanes and that voluptuous thing called barometer!
‘hudson valley happens to be one of the most bird-trafficked deltas on the planet, I know this of my own experience …. hope you will enjoy this!’
the composition is the second in a larger series of ‘lobby soundscapes’ created with the help of microsoft and sister city. the first was a piece from experimental artist julianna barwick that launched alongside the hotel itself in april 2019.
project info
artist: björk
company: microsoft
project: kórsafn, lobby soundscapes
location: sister city, new york
ARTIFICIAL INTELLIGENCE(92)
Sep 29, 2020
Sep 28, 2020
Sep 11, 2020
Aug 31, 2020
Aug 13, 2020PRODUCT LIBRARY
Oct 22, 2020
Oct 12, 2020
Oct 09, 2020
Oct 05, 2020
Seems like the Weather Warlock’s http://weatherfortheblind.org/ has been reimagined for a corporate client.
You can’t say it’s not AI without a cogent argument of what is and what is not AI. Current AI is mostly a form of Narrow AI using deep neural networks that were designed to act like the organic form. It is very likely a form of one of these networks called a GAN (Generative Adversarial Network) that encodes and learns the meaning of input data and then mimics the learned data and then there is an ‘adversary’ that provides the reinforcement of whether it ‘sounds’ like bjorks music.
So you have two AI’s one that can recognize the difference between bjork and not-bjork and then you have the generator that keeps trying to generate things that sound like bjork and uses the feedback of the adversary. Once trained you can then have this generator create new original pieces in the ‘style’ of bjork based on some sort of input, in their case a visual recognition AI that is taking the semantic meanings of various clouds and things it’s been trained on and then generating them in the style of bjork. So what you’re hearing is a form of style transfer from clouds or birds to bjork music…
This isn’t AI. It’s just procedurally generated music based on her catalog. Where does it learn? Where is the reinforcement? Having a lot of random inputs to make something that sounds like human generated music does not make it AI. Training it to identify fluffy clouds is a form of AI, but what do those fluffy clouds have to do with the output? Fluffy or not it will still generate music. Will fluffy clouds mean happier sounding music? How does it know what happier sounding music is and within the context of the guests?