It seems that everywhere we look in the headlines, the intersection of artificial intelligence and the arts is presented in a dystopian light. Artists and creators of all mediums are understandably worried, with the popularity of projects like Dall-E revealing just how easy it is for the average user to generate hyper-realistic or intricately abstract still images by inputting a single sentence into the AI software. The writing on the wall is obvious:, of course the most cynical of corporate executives and marketing teams will try to cut costs by utilizing AI-generated content rather than hiring human designers.

Music and audio professionals are no exception to the fear of an AI dystopia. Just this past Summer, we saw the rise and immediate fall of a supposedly AI-generated rapper named FN Meka. The team behind the digital avatar were dropped by Capitol Records in August, less than a month after signing them, when videos and images surfaced of the character saying the n-word and getting beat up by cops in what can only be described as an instance of digital minstrelsy. The very real humans behind FN Meka’s music, including the vocalist, are white.

The bleakness at the intersection of AI and mainstream music and audio is seemingly endless. Adobe’s Shasta project seeks to automate vocal transcription, mixing, and mastering to the point where the engineer behind the boards may not be necessary for projects that need quick turnarounds. Spotify’s bottom line of data acquisition has been argued as merely a way of generating new tracks in-house to shoot to the top of the most popular playlists.

This is nothing new, however. Capitalism rewards efficient cheapness over genuine artistry in all creative fields. That doesn’t mean that new methods of efficiency in commercial sound can’t be used to push the aesthetics of generative music forward when given to the right group of artists. Every era of music has had its fair share of forgettable albums that utilized the latest in cutting-edge technology. And yet, synthesizers didn’t replace acoustic instruments. Arpeggiators and sequencers didn’t replace gestural electronic instruments. DJs didn’t replace live bands. New developments in audio technology efficiencies have always still required a human being, an artist, overseeing the engine in order to make something with real longevity. We should be able to get excited by new creative possibilities through generative music technology while knowing there will always be soulless crap to sift through. But why take my word for it when we could hear from some experts in the field?

“The line between what can make something good or bad is thin”, says creative technologist, educator, and composer of livecoded generative music Melody Loveless.

“The intention of those developing and using these technologies matter a lot… What makes something a ‘poison’ versus an ‘antidote’ can depend on many factors like dosage - where too much or too little of a drug can make a difference between life and death. The same laptops that accompany us in some of our most intimate spaces and times have a history rooted in war and destruction. When I think about ways to remain positive, I think about the people out there with good intentions. I like to believe in a future where understanding (even as far as literacy) of algorithms and technologies is more commonplace - and with that will come more mindfulness when it comes using said technologies.”

Multimedia artist, educator, and co-host of the Interdependence podcast Mat Dryhurst has his ear directly to the ground regarding AI-generated music. He’s certainly optimistic about the potential musical possibilities.

“It's a wild time. Each of these areas offer very new and alien ways to make art. We got to grow up hearing of tape music, early synthesis and sampling scenes, and from what ,in retrospect, were quite rudimentary tools, we saw a whole new paradigm of art and ritual emerge. This is our moment, and there is no scenario I can imagine in which these powerful new art tools do not produce equally dramatic shifts in culture. There is a lot of AI theater out there, as unfortunately public education on these matters is still quite poor, and there is a lot of work to be done to reorient our thinking into making the industries and habits of the future work out for as many people as possible, and [making sure] that alien new habits are allowed to blossom.”

Composer, educator, and computational arts PhD candidate Matthew Gantt is already imagining a sonic future in which AI can be used as a tool of immersive audio generation in physical spaces well beyond the standard stereo mix.

“If we follow the historic atomization from album to single to algorithmically generated playlists, I wonder if a subsequent development might be ‘sound spaces’ that change based on user preference, input, data, etc. over time (ie: sonic ‘smart rooms’, with all the creative and terrifying dystopian implications intact)... but my hunch is that looking at these things in longview and larger context can reveal more creative possibilities than the typical ‘hashtag disruption’ technosphere conversations would let on.”

If there’s one takeaway from this glimpse into the tech-driven musical future, it’s that technological literacy in the face of seemingly unknown territories of generative sound must be prioritized in all music learning institutions. There are ways we can learn to incorporate generative art into our traditional live music practices, it simply requires the right teachers to present it in a less alienating light. If an educator isn’t even willing to consider DJs, producers, and other music technologists as actual musicians in the first place, how can we expect them to even comprehend the next stages of generative music tech? Fear of the unknown is nothing new in the arts, it’s on the next generation of artists and educators to turn that fear into optimism and excitement.

['Music' made with SampleBrain]