60
This wasn ’ t the first time someone had attempted to imagine what Beethoven ’ s unfinished symphony might have sounded like . In 1988 , the English musicologist Barry Cooper , known for his work on Beethoven , put together his interpretation of the first movement . More recently , in September 2021 , the Swiss orchestra Nexus performed an AI-created , four-minute extract of yet another interpretation . Because it was an artificial neural network ( ANN ) that “ composed ” the piece , they called it the BeethovANN Symphony 10.1 .
Music isn ’ t the only form of art attempted by machine learning ( ML )— a form of AI that attempts to replicate intelligent human-like behavior in machines . In September 2020 , the Guardian published an essay titled , “ A robot wrote this entire article . Are you scared yet , human ?” It was written by GPT-3 , a powerful new AI language generator , which was provided instructions and a 25-word intro . The resulting article , put together by an editor from eight different outputs generated by the AI , led to much excitement about whether AI would make writers and journalists redundant .
Even the theater has “ collaborated ” with AI . In August 2021 , a play was co-created by humans and GPT-3 in London ’ s Young Vic theater . An example from the visual art world is the AI-created piece titled “ Edmond de Belamy ,” the first to go under the hammer at Christie ’ s in New York in 2018 . It sold for almost half a million dollars .
Of course , this isn ’ t an exhaustive list .
AI-produced “ Edmond de Belamy , from La Famille de Belamy ” sold in seven minutes .
THE HUMAN ELEMENT OF AI To imply that AI is capable of creativity or that it can understand and create art seems like an exercise in mixing oil and water , at least in terms of how we imagine both AI and art . Yet , from the above examples , it appears that ML is capable of human-like “ imagination .” If a machine can replace Beethoven , how long before it makes all artists , musicians , poets and playwrights redundant ?
Fortunately , the question itself is redundant . The element that makes AI appear to think and create is actually the humans — the artists and programmers — behind it . Art and its appreciation might be an entirely subjective representation and interpretation of beauty and meaning , but AI isn ’ t just the result of a machine ’ s impartial logic . These so-called intelligent systems come with the perspectives and information they were built and programmed with .
“ Modern AI and modern ML is all about just taking small local patterns and replicating them ,” says Matthew Guzdial , a researcher in machine learning and creativity . He was talking about Beethoven ’ s unfinished symphony on Scientific American ’ s 60-Second Science podcast . “ And it ’ s up to a human to then take what the AI outputs and find the genius ... The genius wasn ’ t in the AI . The genius was in the human who was doing the selection .”
In August 2021 , Evgeniya Fedoseeva , founder and CEO of the knowledge management startup Generation KM and a researcher in AI / ML ethics , attended the screening of the play “ AI .” The production was directed by Jennifer Tang and developed by Chinonyerem Odimba , Nina Segal and — of course — GPT-3 . It was set up as a live workshop spread over three days , with the director , actors , production team and the AI working together to generate the dialogue for the script . “ It was almost like you were viewing a live production of the play ,” Fedoseeva says , the play being cowritten with a “ fair contribution ” from the AI algorithm . Screenwriters would pose a question and GPT-3 answered back . At the end of
PHOTO COURTESY OF CHRISTIE ’ S IMAGES LTD . 2022