Week 8 Lifestream Summary

Kevin Slavin’s TED Talk on algorithms is fascinating and I had never imagined that algorithms could have a physical impact on our environment. Blasting through rock to install cables so that information can travel nano seconds faster worryingly proves that Jarre’s photo of the Himalayas reshaped to echo the vicissitudes of the Dow Jones is no longer a metaphor but a prophecy.

As Slavin points out we’re writing something we are no longer able to read and now have algorithms locked in loops with each other with no control over what they are doing (think Flash Crash 2010).

It was fun playing with algorithms this week and I created a timeline on Tiki-Toki to show the different areas I explored. One of the most interesting things I noticed was with the Youtube algorithm. When I was logged in to my Google+ account the recommended videos were related to videos that I had previously viewed. By logging out of Google+ the options I were given were mainly in Mandarin and clearly linked to my location. In a way I preferred to have the results that related to me personally in comparison to those in Mandarin, but if I were living in an English speaking country (or somewhere like Italy or France, where I speak the language and understand the culture) I would probably feel that the personalised results were too limiting and that I was caught in a ‘you loop’.

There are definitely issues at stake here (as for example with Amazon). What if you order a book dealing with a difficult event in your life (illness, grief) and are then recommended further books on a subject you don’t particularly want to be constantly reminded about? There is also the issue that being in ‘you loop’ may start to impact on your identity and the internet’s distorted picture of us actually creates who we are.

Another thing that really surprised me with Youtube was the comments I was presented with. I had a look at a TED Talk and when I was logged in the comments were better written and supportive of the talk, while when I was logged out they were more critical, contentious and used bad language. (Nice to think though that my algorithm knows that I don’t appreciate the use of swearing in forum comments!)

Similar to Youtube, Coursera recommends courses similar to courses I’ve already done but clearly the difference here is while you may like to watch similar videos, it’s fairly unlikely that you are going to do several courses on the same topic. Here, and in digital education in general, the ‘you loop’ would be very restrictive, limiting the options that are presented to you.

Knox, J. K. (2014). Active algorithms: sociomaterial spaces in the E-learning and Digital Cultures MOOCCampus Virtuales, 3(1): 42-55.

Eynon, R. (2013) The rise of Big Data: what does it mean for education, technology, and media research? Learning, Media and Technology. 237-240. 

4 thoughts on “Week 8 Lifestream Summary”

  1. ‘By logging out of Google+ the options I were given were mainly in Mandarin and clearly linked to my location.’

    Interesting, so it was making assumptions about you based on your location? Yet, Mandarin wouldn’t be the dominant language where you are, is that right? I think there is a really interesting distinction here, that between assuming you are a particular type of person based on your location, or looking at your previous activity to determine that. Both seem fairly limiting?

    ‘the issue that being in ‘youloop’ may start to impact on your identity and the internet’s distorted picture of us actually creates who we are’

    Yes, a central point here, and the one that seems to have significant implications for education. I wonder, from the recommended videos you received, did you feel the combination implied a particular kind of person? (e.g. professional might be lots of work-related stuff, music videos or other entertainment might imply something else). And if so, did you feel any affinity with that identity? Might this be related to the idea that we need to classify students into ‘learner types’ based on their activity data?

    ‘when I was logged in the comments were better written and supportive of the talk, while when I was logged out they were more critical, contentious and used bad language.’

    Fascinating to hear this example, as I’ve never noticed that much of a difference when I’ve tired this, although admittedly not very often. I wonder, are there specific things you have done previously that might have contributed to this censoring? Have you down-voted or reported comments previously?

    ‘it’s fairly unlikely that you are going to do several courses on the same topic.’

    Yes, however you may want to expand your understanding within a certain domain. Indeed, Coursera and edX have both developed specialisations recently, which are sequences of courses that lead to an overall certificate. However, these are predefined rather than being chosen by algorithms, and more nuance would be needed to decide which other courses in the same area would be a relevant step forward. As you say, simply redoing exactly the same subject may not be of interest or value to most.

  2. Jeremy, hi, thanks for the feedback.
    Yes, you’re right about the local language, which is Cantonese. As I could read a good number of characters I assumed many of the videos were in Mandarin but my colleague has just confirmed that many are in Cantonese as the characters are traditional not simplified.

    The combination of recommended videos/courses do imply a certain identity but as they only partially represent my interests I don’t really feel much affinity with that particular identity.

    Yes, agree with the clarification regarding the MOOC courses. The metaphor that springs to mind is that if you buy a black woolly jumper online, you would prefer an algorithm gives you other winter items to match your jumper rather than the same jumper in a variety of different colours :)

  3. Clare, I was interested in your comments about Amazon and ‘the difficult event’ in our lives. I think Amazon, the other platforms we all regularly use now and the associated algorithms have a ‘reminiscent’ characteristic that can cut both ways; e.g. tracing our thoughts, interests and preferences over time. Sometimes I look back at my past un-ordered items in the Amazon shopping basket or wish list and wonder what I was thinking at the time. Others, as you mention have to do with seeking guidance on sensitive health or relationship issues, that have come and gone. Others, with previous interests that slipped away that deserve re-exploring; like Winston Churchill’s history tomes that I intend to order and read some day.

    I think this validates your comments on the ‘you loop’ phenomenon, and gets to Jeremy’s point that yes, I do think we need to understand students’ ‘learner types,’ but that makes education ever-more demanding: tailoring curriculum, delivery modes, and assessments .

    Again, brilliant Tiki-Toki. Cheers, PJ

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>