Like many classically trained musicians, I feel ambivalence towards the introduction of AI into the creation of music. It was partly for this reason that I enjoyed Maya Ackerman's nuanced keynote at the AI & Music Creativity conference at Wolfson College, Oxford, this morning. #aimusic #digital-scholarship
@andrewcusworth
Can you summarize some key points for us? I only compose occasionally, but as musician and especially as a music teacher I am very concerned about the future of AI music generation.
@voltronic I think what interested me beyond the broad shape of that was that May specifically spoke to the idea that investors seek industry disruption (ie decimation of traditional music industry) as a means to generate maximum profit, and that musicians finding ways to harness it and join the conversation properly is a route to avoiding that decimation. There was a lot to think about and agree with in the talk, even if I am deeply troubled by the very real commercial risks to musicians.
@voltronic These were, of course, mainly seen through my filters on the talk. It was sponsored by Digital Scholarship at Oxford (the initiative I work for), and I think we are intending to make it available online. I'll post it when we do.
@voltronic Just following on from your message, would you be willing to express some of your concerns - I suspect we share some!
@andrewcusworth
Here are a few off the top of my head.
1. Loss of work for composers and arrangers
2. Training AI composition engines using work taken without permission
3. Decline in average composition skills (AI fixes your poor voicing/voice leading, etc.)
4. Further homogenization of film/TV/pop music
5. Declining market standards for subjectively high quality music (public brain drain)
6. Decline of formal composition study (becoming a dead end career)
@voltronic This a very good and nuanced list and fits my own thinking quite well - though, yours is better articulated. I didn't have my usual annual opportunity to express disquiet on a public platform this year, which I am slightly resenting, as things are moving quickly. I was quite stunned by the extent of use of some of the 'tools' being discussed at the conference. Have you read about the royalties scam that came to light a few days ago?
@andrewcusworth
No, I don't know what you're referring to.
Regarding my list, something that's indirectly related is the use of sampled or physically modeled instruments for TV and film where real musicians could have easily been used.
I'm thinking of the obviously sampled solo cello used in the Game of Thrones theme. Not to mention that I find that composer incredibly lazy. The Westworld theme is pretty much the same thing with a few tonal and rhythmic alterations.
@voltronic It's an interesting one and I'll look for a link. Essentially, someone used generative tools to write songs, and then bots to play them repeatedly on various platforms, netting something like $10m (actual number may be very, very different, as this is something glanced at a few weeks ago) in royalties by doing so. It's fraudulent and the law will do its work, but it's extraordinary to see how 'AI writes my work and then sells it for me' stuff is becoming more and more widespread.
@voltronic I think one of the challenges is that the concerns around AI cut across different (and differently valued) areas: economy, skills, creativity, human attainment, cultural nuance, quality, and so on. Some of these are already things that are under attack in prevailing culture (the idea that there is 'better quality' music caused me an hour long argument with someone a few months ago). It is indeed troubling.
@voltronic Absolutely! I still have ambivalence for the technology itself, but I think the overall attitude was one that emphasised use in human creation (song-writing &c. by actual people, possibilities of teaching people these practices) rather than as a replacement for human creation.