{"id":8883,"date":"2019-05-13T13:12:39","date_gmt":"2019-05-13T17:12:39","guid":{"rendered":"https:\/\/blog.brainstation.io\/?p=8883"},"modified":"2020-05-12T14:03:42","modified_gmt":"2020-05-12T18:03:42","slug":"how-ai-and-machine-learning-are-being-used-to-make-art","status":"publish","type":"post","link":"https:\/\/brainstation.io\/blog\/how-ai-and-machine-learning-are-being-used-to-make-art","title":{"rendered":"How AI and Machine Learning Are Being Used to Make Art"},"content":{"rendered":"

In the age of automation, there was one job that was supposed to be beyond the capacity of any robot: artist.<\/span><\/p>\n

But recently, technological advances have upended that idea, and machine learning and artificial intelligence (AI) are now increasingly being used to create art across all disciplines, from music to film to painting and even literature.<\/span><\/p>\n

If that sounds like the stuff of sci-fi, here are some examples of how AI is shaking up the arts.<\/span><\/p>\n

Metal Machine Music<\/b><\/h3>\n

For her third album, <\/span>Proto<\/span><\/i>, San Francisco experimental composer Holly Herndon decided to enlist the help of an unusual collaborator named Spawn.<\/span><\/p>\n

In fact, Spawn is an AI recording system. After writing and recording a score with an ensemble in her studio, Herndon would feed the results back to Spawn, which can mimic musical ideas. She would then take Spawn\u2019s contributions back to the band and record the songs again. On songs like \u201c<\/span>Eternal<\/span>,\u201d it\u2019s near impossible to differentiate Herndon\u2019s voice from Spawn\u2019s, or even to discern how many voices one is hearing at any given moment.<\/span><\/p>\n

Herndon has said that while most attempts at using AI to make music are motivated by economics \u2013 for instance, Warner Music recently acquired startup <\/span>End<\/span>e<\/span>l<\/span>, an app that generates personalized \u201csoundscapes\u201d \u2013 she wanted to use the technology to enhance the role of the artist, rather than removing the artist completely. The album has been getting <\/span>strong reviews<\/span><\/a>.<\/span><\/p>\n

\u201cI know I\u2019m known as \u2018laptop girl,\u2019 but I\u2019m always asking myself: where does the human performer fit into this? How do we continue to develop without automating us off the stage? This frees us up to be human together,\u201d she <\/span>told The Guardian<\/span>.<\/span><\/p>\n

Herndon is far from the only musician taking an interest in AI.<\/span><\/p>\n

Ambient producer Sevenism has used several tools created as part of Google\u2019s <\/span>Magenta<\/span> to fuel his prolific output, including <\/span>NSynth<\/span> \u2013 a neural network trained on over 300,000 sounds \u2013 and <\/span>Piano Genie<\/span><\/a>. And singer-songwriter Taryn Southern\u2019s 2018 album <\/span>I Am A.I.<\/span><\/i> was produced entirely using four tools: Amper Music, IBM\u2019s Watson Beat, Magenta, and AIVA.<\/span><\/p>\n

U.K. artist Ash Koosha, meanwhile, even introduced a virtual singer named Yona, an \u201cauxiliary human\u201d that uses a text-to-speech process with the goal of ultimately replicating the voice of a pop singer.<\/span><\/p>\n

\u201cMy hypothesis is that singers will become redundant because this machine will be able to convey every range of the human voice \u2013 an anti-pop manifesto of sorts,\u201d he told <\/span>The Fader<\/span>. <\/span><\/p>\n

A Novel Idea<\/b><\/h3>\n

It makes sense that AI and machine learning could have a major impact on technology-imbued disciplines like film and music, but surely an author\u2019s analog life is the exception?<\/span><\/p>\n

Actually, writers are beginning to look into the potential of AI to assist in composing something even as complex as a novel.<\/span><\/p>\n

\"Author

Author Robin Sloan<\/p><\/div>\n

After Robin Sloan received positive critical feedback for his debut novel <\/span>Mr. Penumbra\u2019s 24-Hour Bookstore<\/span><\/i>, he took a different approach for his second book. Using self-created software that finishes his sentences with a keystroke, Sloan\u2019s method was to write a snippet of text, hit tab, and see what the computer suggested should come next.<\/span><\/p>\n

He built out the computer\u2019s database of texts by using old science-fiction magazines at first, before finding their language too limited. He then added work by John Steinbeck, Joan Didion, Philip K. Dick, and others, as well as Johnny Cash\u2019s poetry and various other texts.<\/span><\/p>\n

\u201cI have read some uncounted number of books and words over the years that all went into my brain and stewed together in unknown and unpredictable ways, and then certain things come out,\u201d Sloan told <\/span>The New York Times<\/span>. \u201cThe output can\u2019t be anything but a function of the input.\u201d<\/span><\/p>\n

The \u201cArt\u201d in Artificial Intelligence<\/b><\/h3>\n

In December 2018, Christie\u2019s became the first auction house to present a work of art created by an algorithm for bidding. It sold for $432,500 \u2013 nearly 45 times its high estimate.<\/span><\/p>\n

\"Obvious

Obvious’ AI art<\/p><\/div>\n

The painting \u2013 a portrait of a stocky gentleman in 18<\/span>th<\/span>-century dress \u2013 was created by Paris collective <\/span>Obvious<\/span>. There was thought to be a higher degree of difficulty in creating a human portrait using AI \u2013 since, as opposed to an abstract work or a landscape, people are likely to notice irregularities in a representation of a person \u2013 but that\u2019s part of what intrigued the team.<\/span><\/p>\n

They fed the system with a data set of 15,000 portraits painted between the 14<\/span>th<\/span> century and the 20<\/span>th<\/span>. The algorithm was then composed of two parts: the Generator, which made a new image based on the set, and the Discriminator, which tried to differentiate between the images created by the algorithm and by people.<\/span><\/p>\n

\u201cWe did some work with nudes and landscapes, and we also tried feeding the algorithm sets of works by famous painters,\u201d <\/span>said<\/span><\/a> Hugo Caselles-Dupre of Obvious.<\/span><\/p>\n

\u201cBut we found that portraits provided the best way to illustrate our point, which is that algorithms are able to emulate creativity.\u201d<\/span><\/p>\n

AI experiments in the art world seem poised to continue, and it\u2019s not limited to paintings. New York artist Ben Snell <\/span>sold a sculpture<\/span> that was created with an algorithm (its training data was an archive of more than 1,000 classical sculptures), while Mario Klingemann <\/span>sold<\/span> his AI-created video installation <\/span>Memories of Passersby I <\/span><\/i>for $52,000 earlier this year.<\/span><\/p>\n

Movie Magic<\/b><\/h3>\n

There probably aren\u2019t many Hollywood directors who could wrap a film shoot in a matter of 48 hours, but Benjamin isn\u2019t like most Hollywood directors.<\/span><\/p>\n

Benjamin is an AI system who created a film called <\/span>Zone Out <\/span><\/i>for a two-day AI filmmaking challenge. Starring <\/span>Silicon Valley <\/span><\/i>actor Thomas Middleditch and Elisabeth Gray, the <\/span>six-minute film<\/span> was created with a combination of thousands of hours of old films and green-screen footage of the two actors.<\/span><\/p>\n