When it comes to composition, how do you see ChatGPT? 

Jacobi: It’s an evolving tool. One of the things that feels important to recognize is that the current ChatGPT is limited to, I believe, data drawn from 2021 and earlier. Some of the most current thinking and available data is not yet incorporated into that aggregation. It becomes easy for students and professionals to lose sight of that. 

It could also be easy for students to think of this as the gift that will solve all their organizational problems or the ability to balance the many demands of life. That’s why it is a responsibility that we have as instructors and practitioners in our various fields to not ignore what’s happening; rather we should face AI tools in ways that will be productive and that will also put it in its place so that it doesn’t become a fad or a crutch that students lean on too heavily.

Can artificial intelligence programs like ChatGPT be used in the writing process? And if so, can they be used responsibly? 

Carter: Since ChatGPT and AI are aggregators of information, you really have to know what you’re looking for to know if it has been aggregated well. I think we might get ourselves into a little bit of hot water if we expect ChatGPT to produce a perfect model of a resume or a review or a commentary. 

But if we use it as a way to analyze writing and rhetorical choices, things like organization, development of ideas, source integration, writing style — those kinds of elements of the writing process can be really useful to help teach students, and the community at large, about what kinds of choices writers want to make in their own writing. 

For example, ChatGPT could work well to produce an essay model and then ask students to evaluate and analyze the model. For example, the professor might ask students “What does the introduction look like in this ChatGPT model? What might you do differently? What about the thesis? Is it clear? Is it specific? How might you revise the thesis to better connect with the audience?” ChatGPT produces a model or multiple models of a genre, but they are not perfect models and should be used to analyze and critique rather than held up as perfect examples. 

One of my concerns with ChatGPT is the potential to rely on it too heavily and bypass the learning process. Writing is about regular engagement, regular practice, and being aware of the choices that we need to make in the process of writing. What we don’t want students to do is bypass that learning experience through the writing process and ask ChatGPT to do the work for them. And we don’t want students to think that ChatGPT models are perfect models that should be copied. They are not perfect. 

Is that a big concern that students will try to do that? Will that be a big issue in the next few years? 

Jacobi: I think some students will. For years, some students have chosen to purchase essays that have been for sale online; there are a number of predatory companies that download curricula and then produce essays and assignments that students can purchase. 

This one’s a little bit different though, and part of our responsibility is to stay on top of what this tool offers and help students understand its limitations.

Carter: During the pandemic, there was a rise in the sharing of course materials among students in various apps or other spaces. So, instructors are going to be concerned about whether students are doing the work and the learning that we want them to be doing. Are they learning the skills, or are they going to go to ChatGPT because they are looking for a quick fix? 

How will you approach that in the classroom? 

Jacobi: Our approach begins with the way we’ve scaffolded the composition curriculum to create unique learning opportunities for students. And teachers will have to continue to do this so that they are asking students to work with writing and produce writing in class and demonstrate their learning through the process of writing rather than solely through an end product. We have been doing that for years. 

Too, many folks are choosing to name AI as a tool that we are, as a culture, adapting to and learning about and learning from, so that we approach it from a proactive, curious and generative stance, as opposed to a punitive stance. 

The initial response in early January was really one of fear. If you look at headlines, people were afraid that human words no longer mattered and that jobs were going to be replaced. Then things calmed down. Now people have a more critical view and a more productive view of what the implications are. Things can always go terribly wrong, but we can also find ways to make them go right. So, it’s important to really think about the trajectory that we have for student learning, the production of certain kinds of texts in our classrooms, where and when we ask for those, and how we give credit to students.

What are some of the big ethical quandaries that you see coming up with ChatGPT? 

Jacobi: One really interesting place to think about some of the ethical quandaries of ChatGPT is the generation of sources. Sometimes there have been sources that have just been completely invented. Back in the spring there was a legal case where a lawyer got called out for citing case law that didn’t exist. 

Training students to be critical users of materials and to do their due diligence to make sure a source is one that they feel confident using is an added incentive for students to think about how to actively participate in their learning about writing, so that they don’t get surprised like that lawyer did. 

One of the things that’s interesting is how AI continues to improve as it learns, or at least that’s the goal. What kind of impact do you see ChatGPT having on the way we write? 

Carter: There’s an area of study in rhetoric and composition called genre theory — which is the study of how genres come about, how they evolve over time, and what we look for in genres. Genres of writing are social acts of engagement between people and people develop expectations for what genre conventions should be included in each genre. In the United States, a resume looks a certain way. A LinkedIn profile is going to generally look a certain way. The same thing goes for an Amazon review. Those content features and design elements are called genre conventions. 

I could see ChatGPT informing what genres look like in the future. Take the resume for example. There are lots of examples online of resumes, and ChatGPT is aggregating those examples. Possibly ChatGPT is going to reinforce what a common standard of a resume in the United States would look like. But if ChatGPT is pulling internationally or if more companies are moving to creative forms of a resume, the genre may end up changing much more quickly than they typically would outside of ChatGPT.