Why Use AI for Course Development? · Instructional Design Insiders

Why Use AI for Course Development?

I know there's a lot of controversy out there about using AI for course development, and both sides have their points. The biggest concern I've seen out there is regarding "AI 'slop'," and the people expressing those concerns aren't wrong. There's a lot of slop out there.

As an instructional designer, it's my job to make sure that I'm not producing slop. Everything I publish has my name on it, whether the learner sees it or not. My manager sees it. My team sees it. I see it. I am accountable for everything that I produce, whether I produce it all myself or with the help of AI.

So let me share a bit of my process — which is still evolving as I learn and experience more every day.

Right now, my solid go-to is building with Articulate Rise with a heavy lean into code blocks to build reflection blocks that collect responses from the learner, track them, and provide a printout option at the end of the course. I've also used code blocks to create highly interactive blocks that display and compare information using filters, enable calculations and experimentation with numbers, and enable practice scenarios.

I'm also experimenting with using Claude Design to produce click-through interactive eLearning with the same block options mentioned above. Claude will publish in SCORM compliant format or html if I want a standalone version.

In both cases, I've been using ChatGPT to create my images.

And I love Articulate Rise and the code blocks, but there's definitely something attractive about Claude Design in that I can restructure and make sweeping updates quickly and easily; no drag and drop required. I now have (mostly) the ability to bring the vision I have for my course to reality, where without AI, I'm limited to what I can code and what my budget allows for stock images or artwork.

The assumption out there, though, seems to be that this means I can just churn out lots of content in a short period of time. The truth is, not if I want quality work and take my accountability seriously.

In reality, it looks like this - starting from the development stage (assuming needs analysis has been done and training was determined to be the right solution):

Type a really long prompt and/or have a long conversation with AI to review all aspects of the learning need with the audience details, behavior that needs to change, preferred learning frameworks to follow, and a high level version of "the vision" while asking for additional suggestions I might not have thought about.
Review all of the information recommended by AI, choose what looks relevant and useful, and reject what doesn't apply
Start with an outline - either mine or one generated by AI and then edited by me - and prompt for a storyboard. In my case, I have an entire prompt already dedicated to my 20-Minute Arc process, so I get get a set of arcs on the first pas.
Review and edit the storyboards (if leveraging Rise) OR go straight to Claude Design and begin developing.
From here I edit block by block or page by page. And it's here that I can make the vision a reality. If I need a specific image, I can go to ChatGPT and ask for it (it knows my branding colors and preferred image style). I can ask for an image in HTML with interactive pop-ups to explain (or bring the image into Rise and use the hot spots - but the formatting sometimes isn't ideal). I can ask for an interactive table with real data (including source citations) that can be filtered to answer questions. I can ask for a set of reflection questions.
At any stage, I can ask for an HTML review publish, or I can work in-line to see what the learner will see.
When I'm finished, I just ask for the output I want. But before I do that, I ask for several checks: proofreading (grammar/spelling, etc.), biased language or scenarios (I try to do this as I go along), accessibility. Even with Rise, I can publish to PDF and run it through AI for polish if my coworkers don't have time.

It's not a short process. Step 5 still takes a long time, because I still demand quality. I still need to weed through the AI "fluff" and "hype marketing" language to get to the phrasing that will be best for the learner. I still need to challenge the structure and move things around to make better sense of flow and presentation. And it still takes a LOT of time to find and/or create the images that I need.

AI can be very helpful, and it's opening doors to creativity that have previously been shut.

But I'm still accountable for what I produce, and I take great pride in my work.

AI Partner Framework

AI Partner Framework for IDs

The 20-Minute Arc (coming soon)

0 comments

Why Use AI for Course Development?