Machine Listening is controversial, but these artist-researchers want you to hear it

Machine Listening is controversial, but these artist-researchers want you to hear it

Research into machine listening has made it out of academic journals and into an upcoming performance titled Machine Listening Songbook, as part of Now or Never in Melbourne, which will be presented 31 August at Trades Hall.

Our devices are listening to us. Digital voice assistants, smart speakers and a growing range of related technologies are increasingly able to analyse and respond to our voices.

Scientists and engineers refer to this as ‘machine listening’, though the first widespread use of the term was in computer music. At the same time, just as machines are ‘listening’, they are also beginning to ‘produce’ sound and music through techniques like generative AI.

Machine Listening is also a group of artist-researchers – Dr Joel Stern (RMIT University), Dr Sean Dockray (Monash University) and Associate Professor James Parker (The University of Melbourne) – who present research and ideas through creative works, such as performances and exhibitions.

Their works explore questions like, what does it mean to live in a world surrounded by listening devices, and in which songs can be generated instantly using AI? The collective’s works attempt to bring together their diverse expertise in machine listening through art.

Machine Listening Songbook Screenshot of the software the researchers created and used to compose with text, sound and colour.

Curator and academic Dr Joel Stern explained the Machine Listening Songbook exemplifies a critical and somewhat satirical focus on music and generative AI.

“The Machine Listening Songbook is a suite of AI-generated songs that attempt to trouble our assumptions about AI and produce uncanny outputs which are weirder and stranger, rather than more homogenous,” Stern said.

The performance looks at what it means to make music when prompting an AI, including with music tools like Suno and Udio, which are currently battling lawsuits from major record labels over the use of copyrighted songs to train their AI models.

Using Suno, they created tracks like ‘wilful copyright infringement’ – which rails against the infringement of copyright in a chaotic mash-up of nonsense and fragment words – and ‘fair use’ – which contains AI-generated lines like “Suno’s wholesale theft of the copyrighted recordings threatens the entire music ecosystem and the numerous people it employs”.

“The songs are almost satirical, using quotes from tech CEOs and others who are part of the tech capitalist economy. It pokes fun at the out-of-control thinking that is fueling the AI bubble,” Stern said.

“We use these technologies in a critical way to draw attention to their social impacts, rather than promote their use or inflate their value.”

In addition to research, writing, and artworks, Machine Listening have produced an expanded curriculum; an online library and interview series; numerous on-and-offline events, lectures, performances; and a browser-based instrument for composing with audio and video via text.

The Machine Listening collaboration emerged out of a previous work on Eavesdropping.

Share

Related News

aboriginal flag
torres strait flag

Acknowledgement of Country

RMIT University acknowledges the people of the Woi wurrung and Boon wurrung language groups of the eastern Kulin Nation on whose unceded lands we conduct the business of the University. RMIT University respectfully acknowledges their Ancestors and Elders, past and present. RMIT also acknowledges the Traditional Custodians and their Ancestors of the lands and waters across Australia where we conduct our business - Artwork 'Sentient' by Hollie Johnson, Gunaikurnai and Monero Ngarigo.