Mark and I (Rie) have developed an OER on Personalized Web, which can be accessed using the link here.
In the OER, we have included visualization and Padlet so that you can engage with the content.
Below are the questions to consider as you go through the OER:
- If you were to implement personalized learning in your classroom, would you see yourself using algorithm-based personalization or rule-based personalization, and why?
- Which personalized algorithm between Facebook, Instagram, TikTok and Twitter (X) do you find the least problematic?
- Tell me the time you implemented personalized learning in your classroom. What worked and what didn’t work?
( 1 upvotes and 0 downvotes )
Great work Mark and Rie! I especially enjoyed your comparisons between social medias personal algorithms, however, it honestly made me reflect on how much it has already influenced my personality.
Nonetheless, to answer your question: Tell me the time you implemented personalized learning in your classroom. What worked and what didn’t work?
When I taught CLC 12 (Careers 12), students had to complete the formidable ‘Final Capstone Project’. For those that aren’t familiar, grade 12 students must work on a project that relates to their career goals, research ideas, community projects, and passions. This project is extremely personalized as students can demonstrate their learning and project in almost any way they want. Although it does of course have necessary criteria and a final presentation, the openness of this project is both a curse and a blessing.
Often when students have too many options, they experience option paralysis where they don’t even know where to start. Even though they are presented with ideas and previous student examples, I have noticed that the getting started period is where the majority of students struggle the most.
In contrast, these projects never cease to amaze me. The creativity, uniqueness, passion, and deep personalized learning make their work so special, with many students gaining valuable skills for their futures.
In all, I think personalization can really support interest-based learning, but may effect time based forms of assessment.
Hey Rie, Mark,
Thanks for your OER submission; it was interesting to think about what personalization meant to me in a time where everything seems to be about personalization and individuality. We throw the word around a lot, but don’t tend to think of what it really means to us, let alone what it means for education.
I’m glad you included a few instances of classroom applications. I’m sure we all, as caring educators, strive to personalize learning for students, as that is the current zeitgeist in public ed. I did have a bit of a hard time understanding what the H5P’s branching scenario was meant to be about, but then I think I realized that it was less about the “computer graphics” topic and more about the idea of branching storylines. I’m curious to know how you feel branching storyline modules like this fit into the personalization ecosystem… Do you feel they’re an efficient use of an instructor’s time, or are there other tools that can reach the same results with less prep time for the instructor? Would a conversation with AI toward a particular learning goal lead to the same learning objectives, but with much less prep time for the instructor? Are they more adaptable than a static branching pathway presentation?
Hi Jake, that is a great question!
Usually, these types of branching scenarios take quite a bit of time to develop. In my context (post-secondary), instructors usually hire students or learning designers to create branching scenarios so they can be used as part of the course module.
I also foresee that AI is more adaptable compared to a rule-based branching scenario. It requires less time to develop than a rule-based approach, but the downside is that I have less control over the output.
Thank you for participating! These are all great comments. It is also interesting that Mmeshi is experimenting with algorithm-based personalization with a chatbot, and Nik is experimenting with rule-based personalization.
Both methods have pros and cons, and it also made me wonder: in the future, with algorithm-based personalization, will we have more or less control over the algorithm?
Great OER! When it comes to the type of personalized learning, I lean toward using rule-based personalization in the classroom. As a teacher with targeted learning outcomes, I would prefer having a clear understanding of what content students are engaging with and why certain learning paths are being suggested. As mentioned in your resource, rule-based systems allow educators to design structured “IF-THEN” pathways that connect directly to curriculum goals and assessment outcomes. For example, if certain students are struggling with balancing chemical equations, the system could prompt a video and some guided practice. I believe it would be beneficial for these students to receive the same educational video and practice as other students in order to remove any ambiguity that might come from a system that might offer differential treatment.
Also, I would be cautious about relying on algorithm-based personalization due to issues like hallucination and the lack of transparency in how decisions are made. In a science classroom, where accuracy and conceptual precision are crucial, even small errors in automated feedback could reinforce misconceptions.
I agree with your thoughts about using rule-based personalization in the science and math classroom. A difficulty students have with determining their own online resources is finding the correct difficulty level of the material (e.g., naming organic compounds differs greatly in high school compared to university). I have had many students frustrated and had difficulty answering questions they found online because they require knowledge/skills not in the current curriculum they are taking. However, there is merit to having the algorithm-based personalization, as long as it takes into account certain specifics of the individual (e.g., which province they are in, grade, curricular outcomes, etc.), but privacy would be a major issue with the stored student data that the chatbot would contain.
It appears that one of the key features of personalization would occur through AI chatbots. What do you think the role of the teacher would be if guidance of prompts is no longer necessary and the AI has developed enough flexibility to not require constant fine-tuning?
In my classroom, I’ve implemented personalized learning using MagicSchool AI. This is a platform that allows teachers to make accounts and create Chatbots that students can interact with. For instance, when completing the Romeo & Juliet unit with my English Studies 8 class, I created a Chatbot in MagicSchool that provided students with information about life in Verona, the play’s setting. Students were then instructed to ask the Chatbot (Romeo) about life in Verona for various social classes, women, and young adults seeking to marry. The Chatbot responded in Romeo’s tone, providing more information as it was prompted. Students enjoyed engaging with this, and they were able to instruct the Chatbot on what they wanted to learn about and how the Chatbot presented this information. For instance, if a student was having difficulty understanding what the Chatbot was writing, they could ask it to “simplify” the language, so that the wording was easier to understand. This made learning more personalized, engaging, and exciting for students as they got this information in real-time and were able to interact with the Chatbot as if it were a figure from Shakespeare’s play. However, one difficulty with this was that some students needed to be more specifically instructed on how to prompt the Chatbot at each step of the way. Some students weren’t able to interact with it beyond a few prompts, think of more questions, or understand what it was saying. So, they needed continuous teacher guidance to interact with the Chatbot, limiting its ability to be personalized to all learners.
Overall, it was a positive experience for myself and my students, and allowed me to work independently with the few who needed further assistance. However, there are still limitations when it comes to making these tools appropriate and accessible for diverse learners.
Hey Mmeshi,
Seems like you’re really on the cutting edge of implementing AI tech in your classrooms. Good on you!
I’m curious about how often you manage to put bots like this together and if you’re finding students interactions with them. Do you think students get as much out of interacting with them as they do through student-student interactions or student-teacher interactions? Is this a “more is more” scenario, or you do feel that these interactions take the place of others, and is there anything lost in that? Geuinely curious as I have not yet tried this with any of my science classes, but I could be curious to do so.