What We’re Hearing from our Clients about AI
Authors: Lynsey Duncan and Devra D'Urzo
In the most recent D2L Instructional Design advisory meeting, we talked with our clients about everyone’s favourite topic these days…Generative AI.
From K12 to High Ed to Corporate, we wanted to learn how organizations feel about AI, whether or not they’re using it, and what they might be most excited and/or concerned about with regards to AI tools. Here’s a summary of that conversation.
Current Attitudes
- There’s a lot of excitement and experimentation to understand the tools.
- Exploration has largely been seeking to automate the mundane tasks (like identifying trends and themes within qualitative survey data) to allow more time for creative work and deep thinking (such as how best to solve the challenges raised in the data).
- K12 and Higher Ed institutions, especially, are considering new assessment models and methods to circumvent AI doing the work on behalf of the student.
- Along the topic of assessments, there were many questions around what constitutes demonstration of learning now and what new or different skills learners need to demonstrate (e.g., is summarizing an obsolete skill now?).
- The consensus is that AI outputs should not be taken at face value - they very much require human review and validation.
- Major concerns pertained to academic integrity, intellectual property, and privacy of student/company personal information.
Future Considerations
- How can AI be harnessed to help instructors grade more efficiently?
- Will there come a time when we do not need human validation of output?
- Will AI be able to create personalized learning paths for students based on individual needs and preferences without significant time and energy from teachers/designers?
- Will it be able to facilitate adaptive learning experiences?
- Are there ways it could search course content and organize based on priority for learners, to provide most relevant information first?
- How can AI create more accessible and inclusive content for all online learners?
- Is the emergence of AI shifting the focus of assessment more towards discussion and debate?
- How can we use AI to support instructors with grading, so we can transition away from MC exams to assess higher-order thinking skills more efficiently?
Current Use Cases
- Organizing survey data and surfacing trends from open-ended comments.
- Differentiated instructions for more personalized learning.
- Using it to generate anecdotes and examples.
- Creating quiz questions that are well aligned with objectives.
- Helping learners create a starting point to build from.
- Customize learning materials to reading levels and learning styles.
- Voice-over tools like Speechify are quick, efficient, and less costly than alternatives.
- As a tutor to support more personalized learning.
Comments
-
Thanks for posting. It's a fascinating topic and one that doesn't have clear answers. The question of what constitutes demonstration of learning has always been a fuzzy one in my mind, even before OpenAI was a thing. A part of me wants to believe that OpenAI will facilitate a shift in course curricula to incorporate a more comprehensive and humanist view of learning, including social and emotional learning, metacognition, self-reflection, and the creative process. These are aspects of learning that OpenAI cannot do (at least not yet), I think.
-
Is D2L contemplating a Brightspace chatGPT plugin that could, on command or based on criteria, generate a response and insert it directly into a discussion thread? That could be an interesting use of AI to facilitate learning and exploration of ideas.
-
Thanks for posting. In the previous comment, the poster wonders if AI-based tools (or, at least, generative text) will "facilitate a shift in course curricula to incorporate a more comprehensive and humanist view of learning, including social and emotional learning, metacognition, self-reflection, and the creative process. These are aspects of learning that OpenAI cannot do (at least not yet)": but let's assume that answer it "not yet", what would these tools look like?
One idea I'm thinking about is to have chatGPT engage as a participant in my discussions. For example, what might it add to this conversation, at this point in the thread?
