A few thoughts on ChatGPT and education
A good friend shared a link to this rather startling video: Introduction to AI Prompts for Educators: Using ChatGPT Preview of the AI for Educators Course, KP Education Systems.. It sparked a conversation about education and the impacts it may have on education. I will not bore you with the details…
The more I read about ChatGPT and other powerful AI programs, the more divided my opinion about them becomes. (I have NOT actually tried them.) I ask myself, are these simply tools like spreadsheets or spell checkers or databases that automate tasks and should be used to reduce time spent on mundane, routine tasks so more time can be spent creating and problem-solving? Or are they truly plagiaristic bots that eliminate the need for human thought at all? I see arguments for both views.
For me, plagiarism has never been the result of malicious behavior by students, but by poor assignment creation by teachers. My old admonition is "if you don't ask for creativity and originality, you won't get it." One of my favorite workshops to give was Are You Punishing or Preventing Plagiarism? which was a hands-on extension of the article Plagiarism-Proofing Assignments published in Kappan. (Some workshop resource links are no longer working.) At the heart of my argument was that all assignments need an element of the personally relevant. Otherwise, the best you can hope for is a paraphrasing of others' ideas. AI amplifies the need for personal relevance.
Mike Eisenberg and I addressed general tech's role in information literacy many moons ago. (See Computer Skills for Information Problem Solving.) But of course this was long before ChatGPT and its ilk. But in summary, info lit skills usually include the ability to:
- Articulate the problem and identify the information needed to answer it.
- Know information sources and locate relevant information.
- Select and evaluate the information in those sources.
- Organize, synthesize, and draw supported conclusions from the information.
- Communicate findings and conclusions to others.
- Evaluate the product and process.
I see AI as being very useful for skills 2, 3, 4, and 5, but human input is still essential in steps 1 and 6. I don't know enough about AI to be much more specific. One thing that concerns me is AI's ability to judge the quality/accuracy of the information it finds with all the mis- and dis-information now available online. Even before information became political, we stressed the importance of evaluating its accuracy (authority, timeliness, bias, etc.).
AI exacerbates the need for us as educators to re-think our purpose in asking students to write. Is it to demonstrate the ability to write a 5 paragraph expository essay with standard organization, transitional sentences, thesis, and conclusion regardless of topic? Or is it to help strengthen the learner’s ability to communicate original ideas? As Simek likes to remind us, "Start with the why."
In using AI as a tool for teachers to create learning materials, I say go crazy. (Much of the video linked above covers this.) I wonder just how many teachers actually create their own rubrics, assignments, etc. and how many simply rely on publishers' teacher guide materials. In my district(s), we encouraged sharing teacher-created support materials (lesson plans, etc) as a good use of our Learning Management System. You got four 5th grade teachers all teaching a single learner outcome, there seems to be no reason why each should do their own thing but instead divide the work. And use ChatGPT as a worker too.
I often think about my kids and grandkids and what they will need to know and be able to do in order to keep from being replaced by AI. I doubt my son-in-law as a minister can be replaced. But I worry my son who is a graphic artist (which to me seems the height of creativity) might be at risk. To those poor souls to whom I have offered career advice over the years, I encourage them to not just be good at what they do, but to aspire to lead/manage those who are good at what they do. Or be the programmer of the program that creates programs.
I am looking forward to the era of totally self-driving cars. I would love a kitchen that buys my groceries and cooks meals that are the healthiest for me. Perhaps an entertainment system that accurately suggests or even creates shows I enjoy (instead of spending way too much time scrolling through Netflix and then winding up watching some old movie anyway.) Or a writing tool that writes books or articles just for me. (Goodreads and GoogleNews are heading this direction.) But then I also ask myself if life would not be pretty boring without the attention needed while driving a winding mountain road or discovering a new author writing in a new style. Would AI ever have recommended I go see the Barbie movie - which I thoroughly enjoyed?
What perhaps excites me the most is that this advanced AI stuff will force conversations about what exactly it means to be human. Perhaps our stupidity is what will separate us from AI in the end. Now wouldn’t that be ironic if the only way we can tell an original human work is by its grammatical errors?
Reader Comments (4)
Somewhere back in the first decade of this century, our district decided to subscribe to a then-new plagiarism checker called TurnItIn. Our office was tasked with creating the training program to help teachers learn how to use it, working with the curriculum specialists, especially in English and social studies.
I was not a fan of TurnItIn and similar products (that got worse over time) and in one early meeting, I said to our district's director of social studies something like "any assignment that's easy for a student to cheat on is probably not a valid assessment of what they can do". She did not like me much after that.
That assessment holds true in the context of ChatGPT and other AI-based systems. I've played with several text generators and a few image generators, and I'm certainly not an expert in this field. However, even I could see the results contained many clear indicators that the product was not created by a human. I'm sure these system will improve over time.
For the near term, if teachers ask students to do work that requires drawing from their personal background and experience, they will be able to spot when students are not doing their own work. If they use the same generic essay prompts and term paper assignments from when I was in school (and I saw many of them still in use), I'm not sure I could blame the kids for using AI tools.
I need to generate a post or two on this subject. Maybe ChatGPT can help. :-)
Hi Tim,
I suspect many, if not most, English and Social Studies teacher ever had a personal relevance requirement made of them as students. Hard to teach in a way one has never been taught. I also wonder if the concentration on standards and "factual" knowledge works against more meaningful assignments.
Nice to know that I was not the only unpopular tech guy out there!
Doug
Thanks for the detailed reflection here! I would push back a bit that AI can't perform items 1 & 6 in your list. Students who know how to set up their prompts can certainly have the platforms answer those kinds of questions. I want my learners to think about what this means for them and their work in a world that now integrates these tools into just about every platform.
I want them to have the expertise to evaluate the responses.
I want them to feel free to disagree with AI conclusions.
I want them to be able to verify or refute the claims the tools are making with credible evidence.
To me, this speeds up the goals I've always had for them by doing some of that initial work from the start.
Hi Jon,
Great to see your name in my inbox! Hope you are doing well.
You make some great points in your comment. Hopefully, all teachers will come to recognize AI as a great tool, not just a threat. Talk about "real world preparedness" - using AI well should be at the top of the list.
If you get a chance, read this morning (Sept 11)'s Star Tribune Opinion Page piece on what separates humans and AI. Thoughtful stuff.
Doug