I’m facilitating week 3 of our Digital Engagement course tonight (which is the third course in our Humanizing Digital Learning Micro program) and a big part of what I’m talking about is how to find the content you are looking for. It’s a messy conversation that’s all wrapped up in some of things that I’ve spent lots of time thinking about and things I only kind of understand. I asked some friends who I happen to connect with on social media platforms their opinion, and some patterns are emerging. Let me spell out what I think some of the issues are and then see if I can recognize some of the patterns of response that I’m seeing.
I should add that what I’m looking for here is IN ADDITION to dealing with misinformation and disinformation. I’ve been using Mike Caufield’s work (maybe this version this year?) for years for this, and will continue to do so. Also, he has a book coming out. You’ll probably want that too.
How do I help students return from an algorithm (search engine/AI) with content that works in my class?
This is my problem statement. If I want to make students finding and evaluating things they find online a part of my classroom, then what should I be doing to help them get the things that I think they should be getting? I love a student-centered classroom as much as the next constructivist, but what is my responsibility for helping them get good stuff? What skills/literacies should I be fostering? How much of my classroom time should i be devoting to this? And, specific to this post, does it matter what I call it?
I want students to be able to go and find content online and work with it in class. My experience is that if you just say “hey go find something” you’ll end up with the first 5 results from a google search. And that’s fair. A lack of scaffolding is hardly a student’s fault. With the new AI systems available now, I’m likely to also get some crafted responses from ChatGPT or whatever. What I’m looking to do is improve those responses to make for a richer discussion in my classes, but support students in developing the literacies that they can then apply to whatever else they’re looking for from an algorithm – whether iut’s information about climate change or their diet or their job.
Student-centered classrooms don’t help students learn!
For context, here, I don’t really care. That is, I don’t care about the kind of learning research that is based on an increase in student ‘retention of information’ or an increase of ‘test scores’. This interview with Paul Kirschner is a fun romp through some of the criticisms. It’s pretty far down the list of things that I value. I mean, remembering things is nice, and important for some things, but I’m happy with remembering being a byproduct of my classroom rather than the focus.
My goal in the classroom is to do the best job I can to help students develop some literacies from a broader set of literacies that I think are important. I care more about a kid liking math and thinking it is useful, for instance, than getting the math right on a test. In this case, I care more about a student getting better at working around corporate sponsored content (where relevant) and finding a more diverse sense of voices than them remembering the five laws or something or other.
I have some core literacies that I think all students should probably learn, and a bunch of others that I’m happy to see students develop depending on their interests or needs. This year I’m moving ‘getting stuff from the Internet in effective and ethical ways’ up my list to ‘core literacy’.
note: I’m probably not going to have an assessment for it.
Does it matter if you call it prompt engineering?
I’m going to follow Karen Costa’s perspective on this one…
This, I think, is good advice. Call it what it is, though Colin Simpson makes the important point that if everyone else calls it prompt engineering and I don’t, then participants wont be able to follow up with other research. Further, Dr. Alison H. Melley suggests that the ‘engineering’ language works for her and that I should include all versions to appeal to as many people’s perspectives as possible. I can see that. So we’ll need to include the idea that it is called prompt engineering in some of the literature. I do worry, however, about the implication of the word engineering. And I’m looking for the every day word I’m going to use in my classroom after the concept is introduced.
Why not ‘prompt engineering’?
I think the word prompt might be ok, though Jon Becker’s ‘algorithm tricking’ might be closer to the plain language that Karen Costa is suggesting. It’s the word engineering that worries me.
Engineering, in my mind, is about a process that you design that brings you to a decisions or an answer. It’s about solving problems. The introduction to engineering on Wikipedia (and i can only imagine how much debate went into it) is “Engineering is the practice of using natural science, mathematics, and the engineering design process[1] to solve problems, increase efficiency and productivity, and improve systems. “
But that’s just the definition. You could totally use the word engineering to mean any process that you use, no matter how nuanced, to achieve any result. For me, though, no matter how recursive your design structure in your version engineering, it doesn’t automatically suggest the thing that I really care about, which is how you search using your values. What I’m talking about isn’t working with natural laws and math to get a right answer. It’s working in a system full of corporate influences, dominant normative voices, conflicting researched positions and perspectives that fully require you to decide what you believe or want before you can find an answer.
Imagine setting up a classroom activity where you were interested in students finding information about how to teach math. I can only guess that if you’ve made it this far into the post, the following distinction might be important to you. If not, I appreciate your thoroughness.
Using the prompt engineering approaches discussed in class, find a good resource for teaching 2×2 multiplication to students.
Using the algorithm tricking approaches discussed in class, find a good resource for teaching 2×2 multiplication to students.
I feel like the first one suggests that we’re looking for right answers, that we’re using the algorithmic systems ‘well’, and the second one acknowledges that were manipulating the system to come up with answers that fit our own perspectives.
George Veletsianos suggests that it depends on the conversation you’re trying to spark
And Angela Stockman says
Final thoughts
I’m hoping to help improve my student discovered content activities. Some of that is just about me setting aside the time in the classroom to go over the results and talk about it with them. Some of it, I think, is that I need to do a better job teaching things like algorithm tricking (or prompt engineering). I think what we call things matters. You may not, and that’s cool. But I think it changes what people think success looks like. If I say I’m going to be a good prompt engineer or a good algorithm manipulator, I think those are different goals.
I don’t know what my students are going to need when they leave my classroom and do whatever they’re going to do. I have this strong feeling, however, that I would like support them in understanding the nuances of complex issues in a way that encourages them to apply their values to that complex issue. So much so, in fact, that I just wrote a book about it.
I imagine a student (i’m teaching BEd students this fall) preparing to teach 2×2 multiplication to their grade 5 students three years from now. They turn to the internet to look for ideas, maybe an AI site, maybe teacherspayingteachers or maybe just a search engine. I would love it if their searching process included a consideration of their values at each intersection of the road. Whatever those values may be.
5 thoughts on “Does it matter if we call talking to algorithms ‘Prompt Engineering’?”
Mentions