Just had an idea to write a post talking about this topic. Will compile all of the community’s comments into an article for our Medium blog.
If you don’t want to be directly quoted, please say so. Otherwise I’ll do “First Name, Last Initial” as the quoting format, so there’s still a bit of privacy.
I don’t feel as though algorithms would have been deemed so sinister or disruptive had they originated from a more diverse team of music listeners and engineers or offered further opportunities for the music writers that they have also displaced by way of inviting their nuanced historic knowledge, but as such, they seem to reinforce bias and champion what is already popular (as highlighted by Liz Pelly’s research, amongst other things), as well as stripping songs and album works of their narrative and context by grouping songs together without considering how they relate to one another offline and as part of wider geographical and specific genre-based communities or scenes. However, I would argue that there is still potential for them to add value to music listening, just not in their current form. The trouble is, algorithms and automated recommendations are viewed as a replacement for human recommendation and interaction and as a means of endlessly generating content for advertising ends, rather than a useful accompaniment to human curation where cultural output and commentary are the end destinations.
Something I saw earlier today that talks about the inequality of streaming and how it feeds into their recommendations:
“And the systems that use Big Data to help us discover new music are also likely to reinforce network effects, unless there is a surge in demand for systems that recommend songs that are unpopular and likely to stay that way.”
For me it’s the lack of transparency about what makes up the algorithm.
Without informing the user (listener) or the sum of it’s parts, there is a distinct lack of understanding about how these services categorise items and thus spit out recommendations/results…
Whilst i understand that exposing the components of the algorithm would lead it to a higher chance of being abused for gain, there is also the opportunity for the listeners to provide feedback on the algorithm, thus inherently making it more human driven then developer focused.
I second the point that @andrew.della has made - issues surrounding transparency often dominate the discussion and don’t offer any real type of understanding.
I believe being honest and transparent about how algorithms are utilised - for example telling consumers how their listening patterns affect musical suggestions - would go further in bringing trust into the related services and provide consumers with actionable data to from which they can begin to learn more.
Agree! For me it’s really simple especially with music. Strong algorithms take the fun out of it. If a song is constantly being suggested to me and undoubtedly many of my friends too (looking at you YouTube) it takes away some magic.
Where as the a person suggests something it comes from a place of emotion.
I resist algorithms because I resist a static theory of myself. I know I am not the same agent in the morning as in the night. Any algorithm that tries to analyze me as though I am a unitary identity or prefer an analytically predictable set of transitions would be a pest.
I really crave a space for constructing novel transitions, and for following the research paths and constructions of others. Not in this life for autofilled narratives.
I had a great experience today listening in shuffle mode for an hour or two, favoriting everything that held my interest after 20 seconds or so. This line of material saved to my queue, I can run back through for further investigation. Very pleasant manner of research. My tastes, interests, and qualities of attention were completely different at the end of the hour than they were going in. I could make a habit of this.
Sounds like a fabulous blog post to me too!
Hi all it’s Coop Fortnight these next two weeks in the UK where Coops are sharing their work. We can do a spot of virtual DJ-ing via twitter into the fortnight/co-ops conversation. I’ve kicked off with Burial. Any favourite releases, hot new tracks, that give a great and diverse flavour of the community, please highlight and we can tweet about them! It would also be great for you to tweet as a member of the resonate community, tag @resonatecoop and #coopfortnight, and we can retweet your human curation!
This popped up today. Thought it might be relevant (about ‘algorithmic accountability’)? https://podcasts.apple.com/gb/podcast/water-music/id1454221845#episodeGuid=d233b9c7-59ed-4f42-8350-9ec15b784152
@robthomas @RobertaFidora @joshuajameswalker @andrew.della @richjensen @peter Reminder re curating some of your top tunes into our twitter timeline for #CoopFortnight! (Roberta I will add one of yours v soon)
Thank you. I’ll have a look at picking something out over the weekend too. I also thought of a couple of people who might have strong opinions on releases and be able to share several suggestions on the electronic front so I’m just off to go bother them now.
Yay! gotta be Resonate artists
@terry @RobertaFidora My bad. Looks like I garbled your intention for a curation/coop voice layer in the design and inelegantly posted a chain of tweets a la personal playlist. Happy to delete and pass the individual cuts over to you if preferred.
I was going for diversity of styles. OCNotes and AGF were the only artists
I was familiar with prior to the exercise. (Labor intensive, but I really enjoyed the work!)
It’s great, good to experiment with different ways of doing things, original idea was there to be built on. We can try different methods. It’ll all still be in the #coopsfortnight stream!
Can’t wait for Resonate to have playlisting on site!
