Working with students to develop professional ChatGPT skills
We're launching a basic course on prompt engineering with ChatGPT ASU — it may be our first ChatGPT-focused course
I may be wrong, but I think we will be teaching Arizona State University’s first ever course on using ChatGPT this summer!
The course is focused on prompt engineering for beginners, and is aimed at teaching undergraduates from all disciplines the basics of getting the most out of ChatGPT (and by extension other large language model chatbots).
Importantly, the learning objectives are focused on résumé building at a time when employers are increasingly looking to make ChatGPT-savvy hires.
At this point, I’m not totally sure that “prompt engineering” as a term will continue to capture the skillset around getting the most out of ChatGPT in a professional setting. What is clear though is that, whatever the term of art, employers will increasingly be looking for people who have a proven ability to craft and use prompts in sophisticated ways, and to assess and iterate around the quality of the outputs of ChatGPT.
Even a quick check with Google Trends shows that weekly searches for “prompt engineering” in the US alone have increased nearly 30 times over the past few months.
Naturally, this is a course that substantially leverages the power of ChatGPT—and part of the story here is how ChatGPT is transforming the process of developing and delivering courses at the undergraduate level.
Initially, the initial draft of the course description, the learning objectives, and the assignments, were all developed using ChatGPT.
Within the course, much of the learning will be through working interactively with ChatGPT. This will predominantly be through assignments that simultaneously help develop and test new skills and understanding.
We’re even experimenting with using ChatGPT to administer and evaluate assignments — all using chained prompts that model sophisticated ways of using the platform.
Of course this is all one big experiment at this point. We may find that no students sign up—either because they don’t see the need yet for adding “prompt engineering” to their résumé, or because they think they can develop the necessary skills without attending a class (although I suspect that employees would prefer the validation of a graded class). And I have a horrible suspicion that we’ll need to revamp the course after the summer—such is the pace of innovation at the moment.
But in the meantime, I’m intrigued to see how this goes!
Prompt engineering doesn't sound that different in essence to learning to use computer technology in the first place. I remember the frustrating hours spent staring at the C:/ prompt, wondering how to phrase my HTLM programming terms to get what I needed out of my computer.
Received my first batch of AI generated assignments this summer. The 'plagiarized' content was very abstracted and typified without any real "content" in the sense of explanatory reasoning, specific applications and discussion of ramifications/limitations. What was especially concerning was the plagiarizers' reaction to this assessment in my evaluation of submissions. Many bright young minds out there but I'm afraid that basic use of AI is going to decrease abstract reasoning and the capacity to apply ideas.