In any case, when I started to put together this blog post a couple of weeks ago, I was signed up for a symposium (or mini-conference?) hosted by a notable LMS provider, and of course, AI was the hot topic du jour.  One of the sessions that I attended was hosted by instructional designers, using ChatGPT, to assist in getting started with course development. Especially if that course development came at the 11th hour when an adjunct finds out that they are teaching a course next week! eek!

In this scenario, an adjunct is asked to create an "INTRO101" course that starts in 2 weeks. As an adjunct, you need to create a syllabus, rubric, discussion, lecture notes, and quizzes. It's noted in this case that this course design may be on top of your current full-time work, so time is at a premium (as is probably the case with most adjuncts).  As someone who's worked as an instructional designer, I appreciate the "let's make the best of it" kind of approach to this, but using generative AI (gAI) to ameliorate a busted system is really problematic! It's also instructional design 101! As part of a needs analysis, faculty don't necessarily need a faster/more efficient way to course design, but rather they need the systematic issues to be addressed.  For example:

1) If the course is an INTRO101 course (not an actual course, just a placeholder for a common course, required of majors, and usually taught by many instructors), the adjuncts should not need to scramble to put together course materials for this course.  The readings, activities, assessments, and rubrics for such a course should already exist, and an instructor should walk in and "turn on the lights" metaphorically.  They should be free to adjust the settings and the materials, but ultimately these should be minor tweaks, not major development work!

2) Hiring for upcoming semesters should not be a "shit's going down in two weeks! Let's roll!" kind of moment. Adjuncts should know at least a few months in advance if they are likely to have a course, and they should be compensated for any additional training required, and any prep they do outside of their contractual period.

3) Faculty are hired as subject experts.  Assuming that someone is hired for their specific subject knowledge (and they're not just another warm body), having gAI create outcomes, course outlines, and assignments (the entire skeleton of a course) really takes away from that subject expertise and starts to frame a course in a way that takes away human agency.  Humans fill in the gaps in what the gAI produces but don't necessarily question it. How do I know?  We see this already with publisher-provided course materials.  I've had more than one professor in my time as a learner that basically used the slides from Publisher, and structured the modules around the textbook, and basically used assignments around the text. Throw in products like My___Lab, and you've removed agency from a human but making it convenient to just go with it.

4) Finally, the material that a gAI produces aren't that great. I have downloaded some ChatGPT outcomes and activities for a class I used to teach (future post), but for this one I'll focus on an experiment I ran with ByteAI. I asked it to give me eight learning objectives for a course titled "Artificial Intelligence in Human Education, Development, and Society." This is what it produced: 

If you'll notice, the objectives are basically like fill-in-the-blanks.  The same starting text can be used for any kind of course. As an aside, it would be interesting to see what it does with other course titles, so I might feed ByteAI some course titles from various university courses. Maybe subject to further experimentation.

Next, I asked ByteAI to provide me with three substantive graduate class assignments for a course titled "Artificial Intelligence in Human Education, Development, and Society." This is what it gave me: 

Again, this is very much a magnetic poetry/fill-in-the-blank kind of output IMO. This kind of output can apply to any class.  Just change "Artificial Intelligence in Human Education, Development, and Society" to your course title 🙄

Anyway, that's what I have for session 2 of #etmooc2. If I have the energy, I'll go through the output of ChatGPT to see what it thinks about what my Intro to Instructional Design & Educational Technology should be ;-)