johavel/iStock/Getty Images Plus
Where does the thinking happen in your class? That is the question that we need to be asking as we professors confront the chaos unleashed by ChatGPT and other AI text generators. The goal of higher education is to encourage students to develop their intellectual tool kit—their academic skills and their knowledge. That means that students must process the material that they have learned and practice discipline-specific skills. But where the thinking happens will differ from class to class, from discipline to discipline.
Emphasizing the disciplinary nature of thinking is essential as professors respond to ChatGPT. For example, I have heard colleagues in science talk about “writing up” their research. When I first heard the phrase, I was astounded. You’ll rarely hear a humanities professor talk that way, because we have always treated writing as thinking. The distinction matters. It may be that in certain disciplines, the disciplinary thinking happens prior to “writing up” the material, and that the presentation of that material is the conclusion of the process. In such situations, students’ engagement with writing may matter less than how they collected the data, tabulated and processed it, and then came up with their own conclusions. In such cases, writing is treated as presenting thoughts that have been completed in a prior setting.
The problem for the humanities is different. Our disciplines have roots in the humanist tradition, which was framed around, in the words of Renaissance historian Paul Kristeller, teaching students “to write and speak well.” Words forged a person, and the capacity to use words well was a defining feature of being educated. That has meant that, for the humanities, the primary mode to represent thought is textual: articles, books and so on. Moreover, cognitive scientists have shown that writing is an effective mode for encouraging students to process material and internalize it. So, for many humanities professors, it is through writing that the thinking happens. We do not consider writing—whether for ourselves or our students—to be “writing up” prior thinking, but to be at the heart of thinking itself. In the humanities, when we outsource our words, we outsource thinking.
But now students can bypass the writing, and thus the thinking. What does this mean for the humanities? Does the next generation, who will grow up in a world of AI, no longer need to learn “to write and speak well”? If so, one of the primary goals of a humanities education could become outdated. Yet, at the same time, students certainly still need to learn how to read well, to criticize texts of all kinds and to develop original thoughts. Moreover, the importance of learning the subject matter offered by the humanities is no less essential today than it was before.
All that said, however, with new technologies available today, we will have to find new ways to encourage students to learn to use words well. Perhaps, for example, the ways we use words will evolve. Speaking may matter more than writing. The humanities seminar, where students parse texts closely and then mine them for meaning with each other and their professor, may be where the most important thinking will happen. This would mean, of course, the end of large lecture courses, since seminars are by their nature small and intimate. Will colleges and universities invest the resources to replace large general education courses with small general education seminars? They will probably need to. The development of AI makes it necessary to invest more resources in the humanities, not less.
Another option is for us as instructors to encourage more in-class writing. We write for many reasons. One reason is to communicate, and the formal academic essay is primarily in this rhetorical mode. It is intended to communicate ideas to another reader. The formal academic essay still has its place, but perhaps it will play a less prominent role in the future relative to other forms of writing. It’s possible that smaller forms of writing—memos, emails, letters to the editor—are better preparation for the writing that students will be expected to do as citizens and in the workforce.
Moreover, we do not write only to communicate. We also write to process our own ideas. As the historian Lynn Hunt has written, “Your thoughts will remain stalled in the fog of infinite possibilities until you start writing them.” That is why we in the humanities consider writing as thinking, rather than simply “writing up.” It is through words that we teach people to construct thoughts. Yet, in a world of AI, that may mean professors need to encourage time in their classes for more reflective writing. We may want to teach students to use words for themselves in more personal ways than the traditional academic essay allows.
Ultimately, academe will probably require discipline-specific responses to how and when it is appropriate to allow a computer to generate words, whether for students or for professors conducting their own research. It depends on where the thinking happens. However various disciplines decide to answer that question, we don’t want students to lose control of the words that go after their name. Whether they are “writing up” their research or writing to think, whether in chemistry or history, students need to learn how to ensure that the words that they create or that are created by computers for them accurately express the ideas that they wish to convey.
Every word we use has subtle meanings and intonations. When ChatGPT uses a word, is it the right word? Does it express what the author wishes to claim? Does it capture the author’s mood? In some disciplines, that will mean teaching students to edit texts that are “written up” by AI generators. But in other disciplines, like the humanities, it will mean developing a deeper relationship with words.
It is easy to panic in the face of ChatGPT. For those of us in the humanities, it is easy to become despondent. But we cannot. In a world in which we may be bombarded by computer-generated text, it becomes more necessary than ever that we teach citizens to read deeply and to analyze texts critically. The question may not be whether but where ChatGPT belongs. And the answer to that question may depend on (re)discovering where the thinking happens in each discipline.