Algorithmic play artefact : teaching@digital podcast:

My algorithmic artefact is hosted on Podbean. There is the podcast itself but also more details of my algorithmic play experience framing this episode as text and screenshots. Would love you comment here or on the Podbean site, cheers:

One thing that has stood out for me in the efforts of platforms to personalise our experience by means of recommender algorithms is how this cultural turn influences what is expected of the services of educational institutions:

“the predefined ‘needs’ of the learner begin to provide the core justifications for education, the role of the educational institution and its teachers becomes merely responsive, one in which the institution exists to supply educational ‘services’ in response to learner demand.” (Knox, Williamson, Bayne, 2020)

In the same way that these ubiquitous platforms become more successful, the more responsive they are to our whims and desires, there is a feeling that for education to “survive” or merely become better, it must be adaptive to what ‘learners’ want.

The assumption being that autonomous ‘learners’ already exist complete as they are, rather than having their learning aspirations shaped by their teachers.

9 Replies to “Algorithmic play artefact : teaching@digital podcast:”

  1. I really enjoyed the podcast and am still in shock that there were over 100 songs with the word algorithm in them! There’s something somewhat reassuring to me (not sure exactly why) about the fact that, as you had been using the service a long time, you couldn’t just change your profile tastes at the snap of your fingers. Perhaps my feeling of comfort around this has something to do with cyber security and that it would be noted if you made wholesale changes, a bit like when the bank notices spending out of the ordinary?

    1. That’s a great point Sean and I hadn’t even considered the cybersecurity angle. And that 100 is only the ones I listened to, I suspect there are hundred more that I just couldn’t be bothered chasing up.

      I made the dubious choice late last night of searching for tracks with the title COVID-19. That was an eye-opener. ” lofi hip-hop COVID-19 tracks to quarantine to” is a new favourite. But someone is also creating an album of experimental rock which is responding to each major ‘milestone’ in the pandemic. It is a strange and wonderful world.

  2. Fantastic artefact, David!

    Really enjoyed both the podcast (some great sounds there!) and the text/screenshot commentary – really insightful. I’d played around with SoundCloud very briefly at the beginning of the block, and glad that you’ve provided such detail framed around Jeremy’s questions. It was also great to be able to engage with your findings and discussion through different mediums!

    It’s really interesting how you comment on the potential privileging of a ‘commercial ethic’ in the algorithmic systems of SoundCloud and others, perhaps shaped by other people, and how perhaps SoundCloud might lose out commercially to others such as Spotify due to its lack of immediacy. You also mention how this might all influence algorithmic cultures in education, and I picked up these familiar commercial/competitive models in Coursera too.

    > ‘One thing that has stood out for me in the efforts of platforms to personalise our experience by means of recommender algorithms is how this cultural turn influences what is expected of the services of educational institutions’

    Absolutely – this is something that I felt while exploring the Coursera recommendation algorithms, tweaking my profile, and also glancing through the privacy policy. The language there was revealing – where ‘Content Providers’ (presumably academic institutions?) make ‘Content Offerings’ to ‘learners’. Presumably, activity on the site from myself and other ‘learners’ will have some influence on what is offer in future, or what is considered commercially viable.

    > ‘This may also mark the impact of algorithmic cultures on education. The expectation of immediate adaptation, flexibility and personalisation.’

    This is a really interesting point, and again one which I was reflecting on looking at Coursera. Thinking about how the expectations and assumptions of myself and others might have been shaped by Coursera’s algorithmic systems – and how this in turn may have been influenced by commercial algorithmic systems outside of education – does make me feel a little uneasy!

    Great work – really enjoyed it!

  3. Hi David,
    I didn’t think there would be so many songs with the word algorithm in it. It feels like really random music content. The real question is, did they use the word correctly or did it just fit in with the lyrics?
    I had a similar experience in mine about Pinterest. I have been using it for a long time so it actually took a bit of effort to get my feed to change.
    That’s an interesting point about how AI is a term that is being used to so broadly now, to improve marketing of a product or just make it sound better in some regard.
    Capitalism dying because w all die does sound fairly possible. At least is what the media is saying is true. Which is probably isn’t, or maybe it is. Anyway, go wash your hands!
    Back to algorithms, that was another great point of relating algorithm play on platforms to social groups. Platforms such as Pinterest or Spotify, it’s about who you know, not the work itself.
    Overall, very interesting and I liked the Podcast style.

  4. Great podcast, David! I have never encountered such pop-ups before – is it how algorithms protect themselves from other algorithms?

    It’s an interesting point that most tagging is still performed by people and the current autonomy of AI is slightly overblown. At my work, we also use algorithms and machine learning for some simple operations, and I have to admit, they are primitive and have a long way to go to understand context and people’s intents.

    Perhaps the most prominent conclusion of yours is that the quality of content is secondary. Rather, it’s the network that defines your popularity on the web. This is indeed, something that is observed in the offline world too. Maybe, it can even be related to IQ and EQ, where the latter proves to be more important for your success.

    Thanks for all these ideas! And the lovely soundtrack…

  5. Great work David and very interesting choice of platform algorithm.
    The fact that there is still a human element involved does question how precise the results can be, especially since these “have a certain set of ideas about what constitutes each mood or genre.”. Netflix does the same thing. It uses people to categorize movie types.

    I also found several of your views very relevant.

    “Students may find it very difficult to make their way if want to change fields of study and also seek out interdisciplinary knowledge in higher education research. ”

    This is one of my concerns about the use of algorithms in MOOCs. Would algorithms suggest the same type of courses? What would happen if you fail a course, would it algorithms suggest a similar one?

    I also agree to the fact that ” It seems that the rise of algorithmically driven streaming services accelerates and empowers of the commercial ethic in the music industry.” It does place more power where there is money. On the other hand, it also offers the rest an opportunity to upload their work somewhere.

    “The expectation that algorithms respond immediately to our changing moods and seeming “choices”.” Is this necessarily a good idea. As humans we often make choices on how we are feeling (which is also important) but does it necessarily mean that the choices are logical. Can algorithms be fooled by our moods?

    Your podcast artefact has generated a lot of queries…as it should. Thank you David.

  6. Hi David,
    I really liked this piece and the podcast was well made and easy to listen to. Our classmates have already commented on the main points that struck me too, which were the amount of songs with algorithm in the title, your findings that after a week of pumping in a new ‘likes’ list, the recommender system was slow to respond, and the point that most tagging is performed by people.

    I didn’t catch the name of your co-host in the podcast I resonated with her statement “the algorithimic landscape is mirroring the human social landscape”. The recommender system is recommending songs to you not on their quality as such, but on the strength of connections of new songs have to other data points.
    As your co-host says, “It’s not what you know it’s who you know. It doesn’t matter how talented you are, you need a network to be introduced to the right people for the right job.” In the same way, the links to the other algorithmic songs (not picked by you) need better connections in order to be introduced to your recommendations playlist.


    1. Thanks for highlighting those points Adrienne. I still wonder whether the social network principle is merely reflecting the human social landscape or whether, as you brought out in your artefact, the entanglement and influence of the digital landscape has changed our perception of work, study and capital in more profound ways than we are aware?

Leave a Reply

Your email address will not be published. Required fields are marked *