In a remarkably short time - since it was released by OpenAI on November 30, 2022 - the artificial intelligence-driven writing tool ChatGPT has sent shock waves through the education world. You can enter a prompt and voila–in mere minutes the chatbot produces a reasonably well-written, if voiceless, essay or some other form of writing.
College professors are retooling their courses and reconsidering take-home essay exams and research papers. High school teachers are doing the same, wondering how to allow for the added class time required for students to write by hand, on paper, with a pen, under a watchful eye. Students are gleefully finding ways to use the time- and effort-saving chatbot.
All of this buzz may fade over time, of course. Technology tools for detecting AI-written text are already being deployed. Perhaps teachers can convince students that using a chatbot is akin to cheating or plagiarizing and count on an honor system to limit its use. Students may see that expressing themselves through writing is a powerful way to learn, and only use ChatGPT as a tool for getting started.
And we may find, as with many new technologies, that we will adjust to it. Chatbots that can write will disrupt some jobs but make others more productive. ChatGPT is perhaps on another level as a writing tool compared to things like Grammarly, spellcheck, or AI-corrected word processing, but it’s not a totally unexpected development. Professors, teachers and students will find educationally valuable ways to use the tool. So it may not be a “battle” after all. It may even become our friend.
ChatGPT in the Context of PBL
Project-based learning has some inherent features that make the misuse of ChatGPT less likely, though PBL teachers might need to adjust some of their practices. They may need to coach students more closely in the development of their written products, and teach them how to use bot-generated writing as a starting point for adding their own voice, ideas, and sources. I can also see using ChatGPT as a research tool or “thought partner,” and it’s being touted as useful for a variety of instructional purposes.
A well-designed project that meets the criteria for High-Quality PBL, taught using effective project-based teaching practices, would not be as vulnerable to “cheating” with ChatGPT as many school assignments might be. A high-quality project should not simply be about conveying information that could easily be created by a chatbot. PBL is a more complex, extended process of inquiry that includes authentic products for authentic purposes and audiences. In a good project, for example, you would find:
Many projects do not (or need not) have a traditional essay or written report as a product. Students may instead create an infographic, an oral presentation, a piece of multimedia, or a constructed artifact. Even if students used a chatbot to help with a written product, they would still need to learn the material to create other kinds of products. If they didn’t, their lack of knowledge would be evident.
Public product and frequent formative assessment
In PBL, students do not create a piece of work on their own and just turn it in to the teacher. They share work-in-progress, including drafts of written work, not only with the teacher but with peers for feedback. They may also share it with outside mentors, experts, real-world stakeholders, or test audiences. Even if students started by asking a chatbot to write something, they would need to make it their own, share it, and revise it in plain view.
Authentic, specific audience
A chatbot could write for a generic/general audience, but would have a harder time writing for a specific audience. Again, students could start with a bot-generated piece, but would have to revise it substantially to make sure it was appropriate for the particular needs of their audience.
Public product with explanation & defense
When students make their work public in PBL, they should not simply show it without explanation. They should explain–to their audience, end-users, readers, and their peers and teachers–the process they used to complete the work. This allows them to demonstrate their understanding of important concepts and discuss the challenges they overcame. It allows the audience to question students and discuss the work. This aspect of PBL would not be viable if a student could only say, “My process was to enter a question into ChatGPT.”
Sources include outside experts, organizations, interviews, surveys, etc.
If possible, students should reach beyond traditional sources of information when creating products in PBL–which is something a chatbot cannot do when it scans the internet. Students could contact local or distant experts to tap their expertise on a topic or problem. The project might include a partnership with an organization, local businesses, or government agencies, and their input would need to be included in a written product. Students might also gather information by interviewing or surveying peers, parents, or people at school or in the community.
Some educators might see ChatGPT as a problem, and others will see it as an opportunity. Will this new technology improve or hinder students’ learning? The answer to that question depends on what we make of it. Teachers in the past adapted to new technologies that were often seen, at first, as threatening to our concept of good teaching and learning, from television and video to calculators to computers. Even in Ancient Greece, it’s reported that Plato was dubious about this new thing called “writing” which he feared would interfere with students’ ability to remember information in their heads. And look how that turned out…
About the Author:
John Larmer is a project-based learning expert. In his 20 years at the Buck Institute for Education/PBLWorks, he co-developed the model for Gold Standard PBL, authored several books and many blog posts, and contributed to curriculum and professional development. John is now the Senior PBL Advisor at Defined Learning.