In our Culture Crossover series we pick up examples of projects that delightfully bridge the worlds of technology and culture. We’ll be reviewing exhibitions, giving you a heads up on cultural events or talks coming up in the UK and highlighting the best techy art.
To read more instalments of Culture Crossover click here.
Half completed scores haunt the music world: the most famous being Schubert’s two act ‘Unfinished Symphony’. Attempts to finish such works have been conducted for centuries, usually by those close to the work or even personally acquainted with the composer. But some have questioned the intention to construct the remainder of musical masterpieces – suggesting that it’s inappropriate to try to recreate the musical sensibilities of great masters. These endeavours have been met with varying degrees of enthusiasm.
Now, in the AI age, a new idea has been plucked up: using algorithms to complete only partially constructed symphonies. This November, the Prague Philharmonic will perform the third movement of ‘From the Future World’, an AI-created composition that completes the unfinished piano piece by composer Antonin Dvorak, 115 years after his death.
To create the composition, Richard Stiebitz and Filip Humpl, creative directors at the Prague office of global ad agency Wunderm approached AIVA technologies (which stands for Artificial Intelligence Virtual Artist), an AI music startup based in Luxembourg.
AIVA is accustomed to crafting emotional soundtracks for films (unbeknownst to many, most of the music in films is already written by AI, which is cheaper than having a piece performed by a real-life orchestra). This was its first request of this nature.
The composition process took around 72 hours, which included time to retrain AIVA on a database of 30,000 scores, as well as all 115 of Dvořak’s opuses to be able to mimic his style. AIVA produced hundreds of example, of which the team selected those that most closely resembled the style of Dvořák.
However, this isn’t the first attempt to finish a symphony using AI. Early in 2019, Chinese telecoms company Huawei presented a completed version of Unfinished Symphony at London’s Cadogan Hall that was written using AI from its smartphone and the film composer Lucas Cantor. Not everyone appreciated the attempt though. Goetz Richter, associate professor of violin at the Sydney Conservatorium of Music, University of Sydney, denounced the effort as a poor imitation of Schubert’s original style, overly grandiose and two dimensional.
The idea of using AI to imitate a music artist is a theme touched upon in the most recent series of Black Mirror (read our review here), where Miley Cyrus plays a pop star in a coma, whose music is produced by decoding her brainwaves and using technology. If these efforts prove successful in real life, the possibilities are overwhelming: an AI that can paint in the style of Picasso? Or design buildings like Zaha Hadid?
If we envisage art as a conduit for the emotions of the creator – a channel of emotion flowing between creator and appreciator whether expressed through splashes of paint on a canvas, or ponderous notes on a piano. Patently, music created by AI did not come from the same place as it would for human composers, but does this mean that it necessarily rings hollow?
What’s important to note is that all of these attempts relied on a human counterpart to the AI, to curate and hone the material the programme produced. This is also true of the wider AI music space where most algorithmic output is still complemented by a human.