Beethoven Symphony X: AI Stunt or AI Genius?
I recently came across one of the most interesting examples of AI being used in music - the Beethoven X AI Project - a documentary film that traces the use of AI to help finish Beethoven’s unfinished 10th Symphony. Believe it or not, this is an “old” story - the premiere of the completed symphony was in October of 2021, but the implications are right at the forefront of many conversations about AI that I have had recently. A company called Playform and a team of machine learning experts and composer used AI to analyze the fragments of musical sketches that Beethoven left behind when he died and then compose a finished version of the Symphony 10 which was then performed by an orchestra. After I visited the site and heard the finished product, I wondered whether this was a stunt or an amazing example of AI in action. I’ll let you decide…
The most obvious question is why use AI to finish the composition? Hasn’t anyone else tried before? To me, the obvious answer is why would anyone try? The effort won’t be as good as if Beethoven finished it. This project is a showcase for AI technology, and I will admit that the end result is pretty good. Here’s a little background information on how this project came to fruition.
When Beethoven died in March 1827, he left behind 40 sketches for a 10th unfinished symphony. A team of experts in machine learning and musicology were gathered to try to use those sketches and AI to finish what the Beethoven never could. There are MANY articles that trace the process, which included training the AI algorithm using the existing 9 symphonies written by Beethoven, as well as garner the reactions of both musicians and audience members. The reaction was mixed, as you would expect. Here is the complete Symphony 10 for you to listen to:
I will be the first to admit, as a non-Beethoven musical expert, that is sounds like Beethoven to me - or at least the late Classical style. Knowing that the algorithm only had tiny musical snippets to go on, the technology is pretty incredible. But what about the human side of the music? The creativity? While I’ll let you jdge that for yourself, I am not bothered by that at all. If this was played at a concert hall, and I didn’t know that Beethoven hadn’t completed the work, I would have no reason to question it. Would I go out and buy the CD? Probably not. Did I feel anything when I listened to it? Not really. In fact, I wrote this post while listening to it the second time, and I considered it basically background music.
The larger conundrum here is what this technology means to composers now and in the future? Can AI compose music? Absolutely? Is it passable? Definitely. Is it great? Meh. Do composers need to consider how to harness the power of AI for their own creative efforts? Maybe. Can we, as a society, stick our heads in the sand and pretend that AI won’t have major implications for our future? No way. Pandora’s Box is open.
All of my thinking over the past 9 months about AI and its meaning in education and the creative economy is how to do everything we can to make it a force for good. A tool for us to use to create amazing new things. While I admit that I am fearful of it being used the wrong way, I firmly believe that as an educator and musician, we need to address it head on and open it up to our students in the most meaningful and inspirational way possible. I’m not sure if I’m a fan of AI or not, but I know that it’s here, it’s not going away, and that it is the most disruptive technology since the Internet.
What do you think? Do you like what AI composed? Is attaching AI to an unfinished work by a genius a good thing or a stunt?