LONDON – During its first 30 seconds, the song Listen to Your Body Choir”Is a rhythmic pop melody, with a female voice singing accompanied by a soft piano.
Then everything begins to fracture, as restless rhythms and samples merge with strange lyrics such as “Are cars with lizards coming?” and a robotic voice intertwines with human sound.
The transition is intended to evoke the co-author of the song:
Imogen Heap, judge of a contest and Max Savage, participant. (via AI Song Contest)
“Listen to Your Body Choir,” which won the AI Song Contest, was produced by MOGII7.ED, a team of musicians, academics, and AI experts based in California.
.Gave instructions to machines to “carry on” the melody and lyrics of “Daisy Bell,” the 1892 Harry Dacre tune that became commonly known, in 1961, as the first song sung using computer voice synthesis.
The result in “Listen to Your Body Choir” is a track that sounds both human and machine-made.
The AI Song Contest, which made its debut last year, is an international competition that explores the use of AI in songwriting.
A panel of judges led by music Imogen Heap and including academics, scientists and songwriters praised “Listen to Your Body Choir” for its “rich and creative use of AI throughout the song.”
The contest received 38 entries from teams and individuals from around the world working at the nexus of music and artificial intelligence, whether in music production, data science, or both.
They used deep learning neural networks —Computer systems that mimic the operations of a human brain — to analyze massive amounts of musical data, identify patterns, and generate rolls, melodies, harmonic progressions, lyrics, and even voices.
The resulting songs included the bewildering 90-second sludge metal thrash punk from Dadabots and Battery-operated steamy electronic dance instrumental, made by a machine powered by 13 years of trance music in 17 days.
While none of the songs are likely to make the Billboard Hot 100 chart, the contest’s lineup offered an intriguing, highly varied, and often bizarre glimpse into the results of human-AI collaboration in songwriting and songwriting. potential for technology to further influence the music industry.
However, the use of AI to create music has not yet fully reached the mainstream, and the song contest also demonstrated the limitations of the technology.
“Robots can’t sing,” said Justin Shave of Australian music and technology company Uncanny Valley, who won the AI Song Contest last year with his dance-pop song “Beautiful the World.”
“I mean, they can, but in the end it sounds like a robotic voice super modified, “he added.
Only a handful of contestants in the AI Song Contest were made up exclusively of raw AI results, which sound clearly warped and muddled, like a glitch remix submerged in water.
In most cases, Artificial Intelligence was limited to proposing song components that were then chosen and performed, or at least perfected, by musicians.
Many of the results would not clash in a playlist between songs created entirely by humans, such as “I Feel the Wires“from AIMCAT, which won the public vote in the contest.
Artificial intelligence becomes very useful when it produces an infinite flow of ideas, some of which a human being would never have considered, for better or for worse.
In a document accompanying his song in the contest, MOGI7.ED described how he worked with technology both as a tool and as a collaborator with his own creative agency.
That approach is what Shave called “the happy accident theorem”.
“You can put some things in an artificial intelligence or a machine learning system and then what comes out really sparks your own creativity,” he said.
“And you say, ‘Oh my gosh, that would never have occurred to me! And then you riff on that idea.”
“We get mad at the machine,” he added, “not at it.”
© 2021 The New York Times