This week’s topic for Stephen Downes’ E-Learning 3.0 MOOC is Data. From the synopsis that Stephen provides for the week we read that…
…. there are two conceptual challenges associated with this topic: first, the shift in our understanding of content from documents to data; and second, the shift in our understanding of data from centralized to decentralized.
The first shift allows us to think of content – and hence, our knowledge – as dynamic, as being updated and adapted in the light of changes and events. The second allows us to think of data – and hence, of our record of that knowledge – as distributed, as being copied and shared and circulated as and when needed around the world.
To try and make sense of this topic I have watched three videos this week.
Personal Learning vs Personalized Learning: What Needs to Happen Oct 24, 2018 Online Learning 2018, Toronto, Ontario, Contact North. This special briefing explores personal learning as the future of learning, explores why it’s important, the tools which enable personal learning and the significant potential of personal learning as a key to life-long learning and the skills agenda. URL: https://www.youtube.com/watch?v=mVnjet3cKfU
This was the video that most resonated with me and related most to my personal interests. What I like about Stephen’s work is that he doesn’t forget to ask the question ‘why’, i.e. the ‘why’ of learning analytics for learners, rather than just the ‘what’ and ‘how’. In this video Stephen tells us that there are two approaches to learning, personalized (formal learning, which accounts for about 20% of our learning) and personal (informal learning, which accounts for the rest). This slide (7) from his presentation ( https://www.downes.ca/cgi-bin/page.cgi?presentation=497 ) provides a clear overview of the differences.
Stephen then considered how we can support an approach which promotes personal learning through discussion of three major themes: choice, ownership and community. In this video Stephen says of learning analytics that it should be for learners so that they can track and understand their own progress. This would mean, in terms of the three major themes, that we can choose what to work on (create our own learning paths), where to store our data and what data to store; that we own all our data and have control over how it is used; and that we are free to work openly and create our own learning communities with whom we can share our data and from whom we draw support. Learning analytics will help us to keep track of our data (which will be distributed over various locations on the web) and self-monitor our personal progress. Personalized learning, whilst still useful and necessary in certain contexts, does not allow for the autonomy necessary for personal learning. The big question raised by Stephen was ‘how can we make this happen?’ i.e. how can personal learning be promoted and recognised in today’s education contexts.
AI in Education Symposium – Introduction: Oct 24, 2018 Artiﬁcial Intelligence and 21st Century Education in Ottawa, my brief introduction and posing of a problem. URL: https://www.youtube.com/watch?v=WENb9N2gnpQ
In this 6-minute video, Stephen introduces the AI in Education Symposium in Toronto. He asks can AI solve the problems of society, since society has now become too complex for its problems to be solved by a few elite, privileged groups? He says that as society gets more complex it becomes increasingly difficult to govern. In the future we will need to teach each other and govern ourselves as a society. We will have to move from a society based on identity, nationalism, religion and language to a society based on consensus and collaborative decision making. The question posed was – Does AI offer us lessons into how to do this? I can see how this is related to the themes developed in the ‘Personal Learning vs Personalized Learning: What Needs to Happen’ video.
Conversation with Shelly Blake-Plock Oct 24, 2018 Week 1 of E-Learning 3.0 with Shelly Blake-Plock, Co-Founder, President and CEO – Yet Analytics. URL: https://www.youtube.com/watch?v=dsmdwnUwKkA
This third video was the E-Learning 3.0 MOOC course video for the week. In this conversation Shelly Blake-Plock described his work in Yet Analytics, a company which focusses exclusively on learning analytics and works with the K-12, corporate and military sectors in the US, to help improve learning content and instruction, and improve the management of data. The system they have developed for tracking learning experiences and performance is known as Experience API (xAPI). Shelly claimed that this system goes beyond how a traditional LMS is able to analyse content and activity. xAPI is able to pull data from the physical world (sensors etc.), mobile devices, games, etc. This data is stored in a secure Learning Record Store, which can then provide automated data visualisations to support learners in understanding their progress.
In watching this third video, it seemed to me that there is a mismatch between Stephen’s aspirations for learner autonomy and the learning analytics systems being developed by Yet Analytics. Questions that were asked by Stephen and others on the course, were:
- How would this work with distributed data (remembering that distributed data allows for choice, ownership and community, as well greater security)?
- Who owns the data/records?
- What are the ethical implications of these developments?
- What are the privacy and governance issues?
- How will the data tell us what learners have learned/understood, as opposed to what they have ‘done’, in terms of number of views, clicks on documents etc.
These are important questions for Yet Analytics to answer if they are really going to provide a system that goes well beyond what a traditional LMS can do and recognises a ‘personal’ learning approach to education.
Finally, as a result of watching these videos and thinking about learning analytics this week, I have wondered what might be the implications of measuring and monitoring everything we do. Is there a danger that it could be taken to excess, such that we treat our bodies like machines, become super-competitive, self-centred and self-absorbed?
Shelly Blake-Plock has pointed out that there are some errors in what I have written about his work, and has responded to the questions listed above. Please see his comment below.
Related blog posts
There have been some interesting posts from other course participants related to all this. See for example:
Geoff Cain – Week 0: Seimens and Downes on AI – http://geoffcain.com/blog/ai/week-0-seimens-and-downes-on-ai/
Roland Legrand – An Experience API for learning everywhere (also in virtual worlds) – https://www.mixedrealities.com/2018/10/25/an-experience-api-for-learning-everywhere-also-in-virtual-worlds/
Matthias Melcher – #EL30 Alien Intelligence AI – https://x28newblog.wordpress.com/2018/10/24/el30-alien-intelligence-ai/
Laura Ritchie – #el30 Notes Week 1 – https://www.lauraritchie.com/2018/10/25/el30-notes-week-1/
Just a few clarifications
— The system they have developed for tracking learning experiences and performance is known as Experience API (xAPI). —
xAPI is not a system and we did not develop it at Yet. xAPI is a data specification and it was developed at the Advanced Distributed Learning Initiative at the U.S. Department of Defense. See the LTSC Tech Report on xAPI Implementation for a more complete description: https://drive.google.com/file/d/1AdMfGl_0-0jKL5cNkik3iEaxiq2SACvf/view
— It seemed to me that there is a mismatch between Stephen’s aspirations for learner autonomy and the learning analytics systems being developed by Yet Analytics. —
I think it’s more a matter of implementation. You can use xAPI to increase learner autonomy. Alternately you can decide to implement it in a way that has nothing to do with learner autonomy. In terms of our own work at Yet, we’ve worked with clients taking either of these (and other) approaches.
— How would this work with distributed data (remembering that distributed data allows for choice, ownership and community, as well greater security)? —
xAPI was developed at Advanced Distributed Learning. It’s designed for distributed data. Note that distributed data does not necessarily have anything inherently to do with choice, ownership, and community — for instance, some large corporations implement distributed learning solely for the purpose of compliance. Again, how a technology is implemented is the thing.
— Who owns the data/records? —
Depends on the implementation. For example, certain types of employee records are often owned by the company on account of IP matters (for example, where training data could be used to reverse engineer proprietary methods).
— What are the ethical implications of these developments? —
Great question. Open floor.
— What are the privacy and governance issues? —
As with any online communication technology — there are privacy, governance, and security issues related to implementation. Some of these issues are technological in nature, others are policy in nature.
— How will the data tell us what learners have learned/understood, as opposed to what they have ‘done’, in terms of number of views, clicks on documents etc. —
Completely depends on implementation. You could track whatever activity you like. If the activity contributes to a digital footprint, you can track and measure it. The question should be: what are you trying to measure and why?
— These are important questions for Yet Analytics to answer if they are really going to provide a system that goes well beyond what a traditional LMS can do and recognises a ‘personal’ learning approach to education. —
Thanks for this post. I apologize in advance that so many of my responses basically say “matters on implementation”, but that’s sort of the nature of almost any data specification.
Hi Shelly – many thanks for taking the time to respond to the questions in this post, and more importantly for correcting my errors, for which I apologise.
Also thanks for posting the link to the report.
Best wishes, Jenny