The cost of automation: How AI will affect the livelihoods of working musicians

This essay is an excerpt from What’s Good: Creativity in the Age of Artificial Intelligence, a zine we’ve co-created with Refraction that focuses on insights, learnings, and creative applications of AI technology.

The promise of the AI hype cycle was that it was going to do all the jobs we didn’t want to do, freeing us from mindless drudgery to embrace our “uniquely” human gift of creativity and knowledge creation.

For many, the vision of replacing drudgery had a lot more to do with spreadsheets than with creative industries. However, recent advances with creative AI have caused equal parts excitement and existential fear as artists grapple with the implications of an increasingly automated future.

Copywriting, graphic design, and music creation would not be the first artistic vocations to feel the effects of automation. Hand-woven rugs are still a high-ticket luxury item today, but those making their living as rug weavers are few and far between. Manufacturing rugs is way faster and cheaper, and that is great for most consumers who just need a rug.

Is this music’s destiny? How close are we already?

Most music consumption is passive in a way where people just need music to play as they are going about their day — working, exercising, driving, cooking etc… Or they just need music to go behind some visual content or ad. In scenarios where tuning out the music is the point, where context and narrative don’t matter, say with focus or sleep music, it’s tempting to feel less mourning for the loss of that work.

In our research at Water & Music, we observed two separate camps of music AI companies emerging, with different operating philosophies around whether music is a creative practice to be augmented, or merely a content production activity to be automated. We have spent a lot of time trying to come up with the best, most succinct terminology to describe these two groups; for now, we will borrow terminology from former Amper Music CEO Drew Silverstein of “artistic” versus “functional” approaches to AI music creation.

→ Artistic music AI tools serve hobbyist and professional artists who are passionate about music creation not as a means to an end, but rather as the end in itself. These companies intend to grow the market of people who self-identify as “musicians,” and who are looking to use AI to upskill themselves musically and push the boundaries of what’s possible creatively. The competitive landscape for this category is more in the realm of DAW, plugin, and sample companies like Ableton, iZotope, and Splice.

→ In contrast, functional music AI tools serve content producers who are looking for cheap, easily accessible, copyright-safe access to music that meets their production needs consistently. These producers include both individuals (e.g. creators on TikTok and Twitch) and groups (e.g. marketing teams at big brands) as customers, and the use cases involved treat music as a commodity — specifically as background fodder for other kinds of content, such as videos and podcasts. The competitive landscape for this category is more in the realm of library music companies like Epidemic Sound.

Interestingly, many music AI startups that start out as “artistic” ultimately pivot to or develop B2B licensing services that are more “functional” (and map to existing, proven business models around production music libraries), in order to subsidize more experimental efforts focused on cutting-edge creator tools.

With this breakdown, there’s a critical evaluation that needs to happen among both artists and platforms when it comes to the commercial impact of AI.

The wider AI vision of automating away less desirable jobs fails to acknowledge the importance of mundane tasks in the tapestry of creative gig work. Every music creative — producer, musician, recording artist, composer, etc. — has multiple products and services in their suite of offerings that they piece together to support themselves financially. Producers may write with and produce songs for their favorite recording artists in the same week that they write themes for podcasts or television shows where audience members don’t care about the credits. They may create sample packs to sell direct-to-consumer with their own unique artist branding, or through larger-scale libraries and aggregators where their own name recognition falls by the wayside.

Every music creative will rank these products differently across creative satisfaction, pricing power, and percentage of income based on their own personal situation; there is no cookie-cutter formula for career success. The one throughline, though, is that creatively satisfying and commercially successful endeavors are rarely the only products that artists offer. Many listeners might be surprised at how important less creatively satisfying jobs are to the businesses of the artists and producers with social metrics and streaming numbers that one might assume were fully paying the bills. In other words, oftentimes the mundane is critical to funding the creative. Less desirable creative jobs can still be critical to artists’ development, representing hours spent on one’s craft and technical skills.

Therefore, I think it’s prudent for creatives (myself included) to take a very intellectually honest look at their product offerings and ask: Which would be automated away first if the arc of this technology continues? What steps can I take to help ensure that I can continue creating for a living? What AI technologies exist today that I can leverage to be more competitive, or that may become a competitor… or both?!

It seems at the very least that the nature of creative work will fundamentally change. Someone who got into music to write music may find themselves in a completely different workflow depending on what music product they are selling. Fortunately, even as technology has advanced, “functional” music AI products are still augmentative rather than completely substitutive, and even those which could technically act as replacements could also easily be leveraged by musicians and composers themselves to improve and streamline their own creative processes. Advertising companies will still need folks to operate music AI technology and work with the generated music in a multimedia context.

This is all very unsexy talk around creation, but that’s the point. Being a working artist often involves a mix of sexy and unsexy, glamorous and pedestrian, inspired and executory. Art and artists will do what they have always done: Evolve.