ChatGPT Series — #2: Where the Rubber Meets the Road: ChatGPT, Academic Writing, and Plagiarism
Over the past few months, we have both facilitated and attended workshops on ChatGPT for university instructors. Some of the major concerns that pop up include assessment (do we need to modify it), plagiarism (how to detect it), and reliability of ChatGPT outputs (can we trust it). The simple answers to these are: we do, you really can’t, and yes-and-no-it-depends. The fun and productive part begins when we turn to address these concerns by revisiting course assignments, rebuilding lesson activities, and rethinking how we teach.
At this point usually a hand goes up and a voice asks: But doesn’t this mean that every course will turn into a course on how to use ChatGPT?
Let’s unfurl this thread a bit before trying to answer the question.
ChatGPT is here to stay. Students are already using it. Sometimes they use it in pretty advanced ways. But there is a difference being able to use a tool and using it in a way that is productive in a discipline. It is unfair to ask students to make that transition all on their own, just as it would be unrealistic to expect students with basic literacy to automatically be able to write academic essays. To mediate the shift from taking advantage of ChatGPT for potential ghost-writing and plagiarism to using it as a productive and ethical tool in academic work, courses will need to include training on how to use ChatGPT in the context of a specific course.
How to integrate ChatGPT?
We are already seeing that ChatGPT can be integrated into courses in pedagogically sound ways. At the Erasmus School of Philosophy, Rolf Viervant scaffolds his assignments by asking students to create prompts for ChatGPT, assess and comment on its outputs, and then reflect on the process, all while engaging with the main topic of the course, Rousseau’s The Social Contract. At the Rotterdam School of Management, Dimitrios Tsekouras has integrated ChatGPT into the course on digital business by requiring students to use ChatGPT when delivering reports (similar to how statistics courses require students to use R or Python) and has reshaped the assignments and the accompanying grading rubric to account for students’ use of ChatGPT in the learning process.
The answer to the initial question — whether this means that every course will turn into a course on how to use ChatGPT — is now coming into focus. And it looks like it begins with a qualified “yes”. Yes, one effective way to address the inexorable presence of ChatGPT in higher education is to integrate it into courses. Maybe not all courses will need to do it. And adjusting course assignments and activities will need to follow investing significant effort into guiding students on how to use ChatGPT in the context of a course.
But, that was the wrong question to begin with.
The question we really should be asking is whether reshaping courses around ChatGPT is helpful for student learning.
Why integrate ChatGPT?
Students’ success in a course depends not only on regurgitating factual content, but on building and demonstrating a certain set of skills: critical analysis, writing in the discipline, drawing connections, synthesizing information, replicating processes, applying processes in different contexts — these skills will depend on the course.
One of the fundamental instructional principles is to teach students what we plan to assess them on. If the final assessment is a written assignment, then receiving instruction in how to effectively write in the discipline is essential to student success. And therefore, it is essential to effective teaching. We may prefer to relegate the teaching of writing to courses specializing in academic writing, so we can focus on the content. Reasons for it vary from concerns about covering the assigned content in a limited amount of time to uncertainty about one’s own ability to teach the skills in addition to the content. But the fact remains that teaching content without teaching the skills is not division of labor, but pedagogically less efficient — and arguably ethically questionable — teaching.
Similar to academic writing, ability to practice academic integrity is another skill the mastery of which is more often assumed than actually taught. Even though it can be folded into teaching, many courses choose to impart principles of academic integrity by referring students to the code of conduct. These official statements typically lack discipline-specific examples, instructive comparisons, or opportunities to practice and receive feedback. In other words, they lack guidance and instruction on how to learn and demonstrate academic integrity.
And yet, many courses that hinge on written assignments and academic integrity continue without providing instruction on either of these skills, but rather assuming that students would acquire them elsewhere or pick them up along the way.
Here is where the ChatGPT rubber meets the teaching road.
ChatGPT is putting us in a position where we can no longer delay adding the teaching of skills to the teaching of content. The latest addition to the list of skills is: academic writing with the help of ChatGPT. This includes assessing ChatGPT outputs and using those outputs in ways that meet academic integrity standards. And that means actual instruction — with assignments, examples, practice, and feedback — in writing, academic integrity, and ChatGPT use.
So, what’s next?
This is a new chapter in teaching and learning. It requires professional growth — and yes, time and effort — on the part of university instructors. But valiant early adopters at universities including ours — like Rolf and Dimitrios — are showing that it can be done. And it can be done well, in ways that empower student learning.
As we proceed, we should give ourselves grace by acknowledging that for a fast-moving phenomenon as ChatGPT and other generative AI, answers, solutions, best practices, and policies will not be instantaneous. This is a marathon, not a sprint. It is also not something that will be possible to address through the efforts of a handful of enthusiastic individuals. So, talking to colleagues, talking to students, attending events and workshops, exchanging samples, exchanging ideas, admitting and sharing mistakes, and brainstorming potential approaches…. all of this will be not only helpful but essential as we collectively build knowledge and skills to incorporate this gargantuan new tool into the ways we teach and learn.