ML in the News: Music and Artificial Intelligence

Above: Andrew Blumenfeld, co-founder of Telepath, and Lily Adelstein, creative project manager, discuss the topic of music and AI.

DJ Mag is in the midst of publishing a series on the future of music and the influence AI is having on the industry. It’s a fascinating look inside one of the many sectors of our economy that is being reimagined with the help of machine learning, and- because of the innately creative quality to this particular industry- it especially highlights questions about the relationship between humans and machines.

The article discusses the significant extent to which AI and machine learning are already part of major tools and products that are used by music professionals every day. Specifically, when it comes to music production and engineering, machine learning can help do everything from automatically detect tempo markers to adjusting levels to even generating new music. iZotope, a leader in the space, has worked for years to develop cutting edge machine learning models that can understand the music it is consuming in order to make these kinds of predictions and recommendations for the user.

So much of music production is considered a creative endeavor, and the application of ML in this space adds a new layer of complexity to that. Is the work any less creative when some chunk of it is being completed by machine, rather than by a human? It’s quite possible that the opposite is more true: that the utilization of machine learning to solve for the most technical aspects of the job, leave for the sound professional a scope of work that is the most creative and difficult for machines to do well. 

In many ways, these kinds of uses for machine learning represent a bigger shift that is likely to occur as this technology is more widely adopted across industries. As we learn new ways to leverage the capacity of machines, we will inevitably need to rethink our expectations of humans. It seems very plausible that this will mean an increasing reliance on people to contribute in ways that require integrating knowledge and experiences that are currently very difficult to digitize and make available to machine learning. The premium placed on people with creative and other more abstract intelligences will likely only increase in that reality.

But there are many other interesting human/machine paradigms that are raised in this article and in this context. For example, once music has been generated by AI, how does one credit the source of that music? Who really created it? Who contributed to it? It may seem like a small question- maybe even one just about vanity- but it really points to something bigger than just who or what should win the Academy Award for best sound mixing. So much of human development- in music and elsewhere- is based on a deep understanding and appreciation of prior development. How does this change when our sources of inspiration are generated by machine and not people? 

Artificial intelligence is already changing every industry and so many aspects of our lives. As it embeds itself more deeply in creative work, we may have to confront more quickly these and other complicated questions about the relationship between man and machine, and the work we do together.