The unofficial version of the iconic event – cancelled due to coronavirus – asked 13 international teams to create a song via machine learning
Back in March, the Eurovision Song Contest was cancelled for the first time in its 64-year history due to coronavirus. In its place came an unofficial AI version, which asked 13 international teams to create a track via machine learning.
Livestreamed on Tuesday (May 12), the event was hosted by Dutch broadcaster VPRO, and saw the UK, France, Germany, Sweden, Belgium, Switzerland, and the Netherlands take part, with Australia – which has been part of the IRL Eurovision since 2015 – taking the title.
The Australian entry was submitted by a group called Uncanny Valley, who said their song was a response to the country’s recent bushfires. Called “Beautiful the World”, the artists created the track using a neural network trained on noises made by koalas, kookaburras, and Tasmanian devils.
Uncanny Valley – which is made up of maths, computer-science, and social-anthropology students, as well as music producers – used artificial intelligence to write the melody and lyrics, while the animal samples were used to craft a synth instrument.
Germany’s Team Dadabots x Portrait XO took second place with a song titled, “I’ll Marry You, Punk Come”. To create their lyrics, the group used AI trained on 1950s acapellas, while the music was generated using a collection of neural networks trained on a variety of music styles.
Most entries used both artificial intelligence alongside human contributions, except for Switzerland’s Team New Piano, who came last with their track, “Painfulwords”. The data scientists behind the group said of their entry: “Faced with the choice between making an accessible song with quite a few human interventions, or experimenting with as much AI as possible and then delivering a worse-sounding song, we chose the latter.”
According to the BBC, the contest’s judging panel were “amazed by the wide range of innovative approaches to using AI to create music”. They said: “Composing a song with AI is hard because you have all the creative challenges that come with song-writing, but you also have to juggle getting the machine learning right. The teams not only pushed the boundaries of their personal creativity, but also gave the audience a look into the exciting future of human-AI musical collaboration.”
Listen to Australia’s winning track above, and Germany’s second place effort below.