Meta AI has introduced AudioCraft, a text-to-music AI model that does an impressive job of composing, arranging, orchestrating, and playing in a wide range of musical styles. Not only can it create or mimic a melody, but it also engineers and produces fully finished work, some of which is already suitable for commercial use – and this is only V1. Imagine where the state-of-the-art will be a year from now.
Your Expertise Matters
You are the world’s foremost expert on the music you like. You are also the ultimate authority on how you spend your attention as well as your entertainment dollars. If a piece of music sounds good to you, it’s good. If it doesn’t, you get to ignore it. And, most importantly, if you can’t tell the difference, there is no difference. These are immutable truths about your personal music consumption.
Before we go deeper, the main purpose of this article is to help frame a specific part of a very complex issue. I’m not going to discuss AI training or ethics here. The following thoughts and discussion starters have been inspired by recently released generative AI music production tools. Your ideas and comments are both welcome and encouraged.
Playing Notes vs. Playing Music
I’ve been a professional composer/producer for over 50 years. (You can see some examples of my work here.) So, with a few thousand projects under my belt, I feel qualified to tell you that there is a huge difference between playing notes and playing music. Any professional musician can play the notes you put on a page. That’s their job. But the music that moves you the most is a magical combination of human emotion, technique, the vibe in the room, and countless other variables that transform notes on a page into music. Under the right circumstances, playing music is truly joyful — and the audience feels what the players feel.
AI may be able to mimic every aspect of a sonic experience, but hits are magic. Said differently (and mixing metaphors), if data were the whole story, every originally produced show on Netflix or Amazon Prime would be a hit.
But not all music is hit music. Nor is it intended to be. A fair amount of client-funded music is production music. That’s the music you hear in the backgrounds of films, videos, video games, and commercial messages of every kind. It’s also the music you might not even notice, such as hold music, music in retail stores and restaurants, and the proverbial “elevator music.”
What About the Music Business?
I can also say, based on my professional experience, that music has almost nothing to do with the “music business.” The vast majority of recording artists are entertainers who use music as just one of their communications channels. Generative AI poses no threat to them. In practice, AI tools will improve their productivity while offering new creative avenues to discover. (Think analog synthesizers, drum machines, or sampling keyboards.) The idea that you won’t go see Taylor Swift because she’s used an AI-coworker to help her create and produce her songs is nonsense. You’ll attend the concerts and listen to the music because you’re a Taylor Swift fan, not because you are (or she is) a music purist.
Clients Are Always Looking to Cut Costs
Production music, aka commercial music, will likely be AI’s victim number one. Production music is rarely the music you’d choose to listen to all by itself (excluding great film scores). There are many reasons for this. First and foremost, production music’s job is to add to an overall experience, not stand alone. Then, it is usually mixed with sound effects and dialog, which tend to take precedence over the production music itself. This is the music that generative AI will replace first. The economics are clear.
Professional Musicians and Singers Will Likely Suffer
While the headlines seem focused on pitting subjective human musical excellence against AI, and feature clickbait about big-name talent and high-profile lawsuits, a palpable issue will be the diminishing number of opportunities for professional musicians (and ultimately singers) to work. There’s a big difference between the speculative world of hit music and blockbuster movie music and the work-a-day world of journeyman musicians and singers. When machines are used to make more music than they already make, there will be less work for people who play instruments and sing in choruses.
Follow the Money
As AI tools get better (and they will do so quickly), the economics of the music industry will dramatically shift. So far, the work product from generative AI music models is not protectable by copyright. The U.S. Copyright Office has offered some guidance on this subject, but it is a hotly debated topic and will remain so for the foreseeable future. Once the ownership and IP issues are well-understood, business will adapt quickly.
If You Ask The Wrong Question…
The title of this article not only asks the wrong question, it improperly frames the issue entirely. The question is not whether AI can create commercially viable tracks in various musical styles. Clearly, it can. Nor is it important (other than for academic reasons) to ask if an AI system can create music as well as or better than a human. AI already creates music better than most humans. It is also unimportant to ask if AI can create works that will inspire us and move us in ways that meet or exceed the works we revere from the greatest artists of all time. Answers to these questions (whether you agree with my answers or not) say nothing about the economic impact such capabilities might have.
Remember, popularity has never been a measure of quality, and quality has never been a measure of popularity.
There’s nothing anyone can do to stop this. These tools are an inevitable result of multimodal AI research. Yes, some researchers are focused on audio and music. But the math and the models are just a part of a much larger economic opportunity. In the end, productivity is the key driver of economic success – and that’s where everyone will face the music.
If you want to learn more about how AI is trained and how it is likely to be used in content production of all kinds, (including music) please sign-up for our free online course Generative AI for Execs.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it.
ABOUT SHELLY PALMER
Shelly Palmer is the Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, a consulting practice that helps Fortune 500 companies with technology, media and marketing. Named LinkedIn’s “Top Voice in Technology,” he covers tech and business for Good Day New York, is a regular commentator on CNN and writes a popular daily business blog. He's a bestselling author, and the creator of the popular, free online course, Generative AI for Execs. Follow @shellypalmer or visit shellypalmer.com.
