By tracking how a student responds to failure within the AI tutor, we can generate a productive struggle score
That's an idea-- using AI to measure germane cognitive load!
By tracking how a student responds to failure within the AI tutor, we can generate a productive struggle score
That's an idea-- using AI to measure germane cognitive load!
By analyzing the digital trace data from student interactions in shared documents and team chats
entails a radical openness to the sharing of data
systematic, personalized reinforcement
Entails explicitly outlined information about what's important to remember-- seems a little far-fetched.
But we no longer live in an age of information scarcity. The lecture is a solution to a problem we no longer have. The challenge for colleges and universities in the twenty-first century is to deliver artisanal-quality learning at an industrial scale. For decades, this has been an impossible dream. Until now.
Logic: The lecture format of learning was in place to "allow one expert to broadcast information"- which can now happen in a multitude of ways (cue "flipped classroom")- should be supplanted with at-scale personalized learning. This is the best way to scale the Socratic ideal, short of direct expert-to-learner instruction.
In the liberal arts, we often critique capitalism’s exploitative systems, yet we reproduce the same patterns in our own knowledge economy. We externalize the costs of learning and call it normal.
Great quote here. What does an anticapitalist course design look like?
And here’s the truly jarring part: many of those same publishers are now selling our work again. This time to AI companies without our consent or compensation. I’ve come to label it as academic fracking: extracting value from our intellectual commons, layer after layer, until nothing of public good remains.
What data are these publishers collecting and selling?
The “fetish” for particular kinds of writing—certain tones, certain sentence structures—reasserted itself, this time through algorithms and bias detection tools.
Resurfacing our baises through AI
Effective use of outcomes entails rubrics.
the professor released a custom GenAI prompt designed to interact with learners and tutor them on topics from the quiz.
Advanced ues case: custom prompt that is focused on the assignment!
This time-saving effort allows the instructor to focus time and effort on interacting with learners.
Optimistic spin: saves time to focus on interaction with learners Cynical spin: GenAI can do my job for me and is putting my very job at stake
However, it is oriented toward the instructor, as the GenAI must consult the instructor for additional materials.
Providing content to instructors on the learners' behalf!
Instructor Proxy. In this mode, the GenAI's response is advocating on behalf of the instructor while interacting with learners.
I'm finding this a little tricky to disentangle from the Instructor Assistant mode, because both are aiming to develop materials for the students. Maybe it's that the Instructor Proxy is focused on output-- creating output that goes directly for the learners, instead of being more instructor-mediated? It isn't that clean.
Instructor Assistant.
This can be the focus of an entire workshop/discussion.
intent and the orientation of the user affects interaction
What is the origin of this model?
GenAI is introducing novel ways for learners to interact not only with their peers and instructors but also with autonomous entities
This is some great framing, because it opens up a new set of questions. * How will interactions between students and instructors change? How might interactions between students and their content change? How might interactions between students and their classmates change? [Is this question even relevant? ]
Then we will need to teach students how to work with AI
This I think is the endpoint for academic tech/ instructional designers, but how does this happen? Does "train the trainer" work better or integrating IDs/AT staff into the curriculum design process?
Talking directly and candidly with her undergraduate students
Love how talking to the students is included in this!
Read: Models for Leading Classroom Discussions
We can potentially adapt this content to read more nicely on GLOW by converting it to a google slides deck.
clicking on the +Group button
potential spot for screencast
Maybe a "Need help? Watch this" sort of thing.
It turns out that by adding your assignments to the assignment area, many of these functions of the course management system are automatically activated
This is the main takeaway from this page, right?
I think reframing this as... Let GLOW do the work for you! or something like that might work.
maybe this goes in Module 1?
I think you covered this when you talked about homepage options.
(as an aside.)
Not clear on what this means. Is it like a secondary learning objective? I think you can probably delete "as an aside"- it's good content!
either by mouse drag-and-drop or by the Move option under an item's triple-dots menu.
Would you like a 10-second screencast to illustrate this?
Imagine, instead, if a student had to...
LOVE THIS
Creating a Pages Front Page (Special Instructions)
This page isn't loading for me. 6/10/24 at 5pm.
If others have this problem, I suggest making this a link to an external resource. The link works for me.
s clear/consistent navigation, a logical layout, and easily-accessible resources:
I'd be willing to make an optional deep-dive video into how a structured home page aligns with UDL principles!
The video [1:58]
May be good to edit captions so that the email reads correctly.
Submit a GLOW Sandbox Request Links to an external site..
Important we're all subscribed to submissions here
This sums up our purpose perfectly!
long with annotations regarding the reasons for the edits.
Seems a little far fetched at scale.
The AI generates a draft of the learning design
How is "learning design" defined here?
It could improve significantly if it were trained on annotated data.
What would this annotated data look like?
The end goal is to develop standards for integrating the ongoing ITS research (and other data-backed research streams) into continuous improvement of AI tutors.
Would these standards be visible to end-users?
How are those behaviors grounded in specific research findings about cognitive load?
Future search: cognitive load and LLM-based tutors
Let’s start with the authors not providing outcomes data anywhere in the 50-page paper
!!!!
The paper proposes evaluation rubrics for five dimensions of generative AI tutors:
Can use this in an individual framework too for evaluating the suitability of the tool.
But ChatGPT and similar programs are, by design, unlimited in what they can “learn” (which is to say, memorize); they are incapable of distinguishing the possible from the impossible.
i.e. incapable of reason. Will that change?
On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information; it seeks not to infer brute correlations among data points but to create explanations.
Concise explanation of how machine learning is different than human learning
critically and systematically examines the cultures
Test annotation here!
ecting some text and clicking the button
Here's a note