To the Editor:

I read Jeff Schatten’s article “Will Artificial Intelligence Kill College Writing” (The Chronicle Review, September 14) with interest and chagrin. Schatten does an admirable job of addressing GPT-3 and, by extension, other writing AIs use by students in higher education. He rightly diagnoses that students will use programs like GPT-3 as a “Homework Machine” to avoid writing papers. He offers straightforward suggestions for limiting students’ ability to use these programs. He recommends, for example, requiring students to write about local issues, use class material, and use clear citations. However, when it comes to why students ought to be dissuaded from using GPT-3, Schatten seems to be less clear. In fact, Schatten appears to believe that in the future, all writing will be AI produced and that such a future is acceptable.

Schatten asks the question, “Given the rapid development of AI, what percentage of college freshmen today will have jobs that require writing at all by the time they graduate? Some who would once have pursued writing-focused careers will find themselves instead managing the inputs and outputs of AI.” Later in the article, he suggests how this might work: “Eventually, we might enter the “if you can’t beat ‘em, join ‘em” phase, in which professors ask students to use AI as a tool and assess their ability to analyze and improve the output.” As an example, Schatten talks about Beethoven’s unfinished 10th symphony. The symphony was fed into an AI program, which then composed multiple endings to the symphony. Scholars sorted through the possible endings, pieced together the best parts, and created something that sounded like Beethoven. Schatten suggests that in the future, writing may follow the same process: Aspiring writers will feed a question into an algorithm that pulls data from across a vast body of knowledge and plunks out an answer. The student then sorts the information, collecting the best parts, and compiling them into a coherent, maybe even eloquent, paper. By this definition, the goal of writing is simply to produce a piece of writing.

Ultimately, though, writing is not just about putting words on a page, even if those words are coherent or even eloquent. Writing is about exploring and communicating ideas in a compelling and clear way. GPT-3 may be able to write sentences, paragraphs, and even full papers, but it is not able to carefully consider an idea in its context and offer thoughtful insights into that idea. Managing the inputs and outputs of an AI program will not teach students to think carefully about their topic, to develop their own ideas; it will not teach them to translate those ideas into coherent prose that a reader can easily understand.

In the 2006 introduction to his classic book, On Writing Well, William Zinsser says, “I don’t know what newer marvels will make writing twice as easy in the next 30 years. But I do know they won’t make writing twice as good. That will still require plain old hard thinking — what E.B. White was doing in his boathouse — and the plain old tools of the English language.” Hard thinking is something GPT-3 cannot do.


Andrew Hedinger
Director of Admissions
Western Seminary
Portland, Ore.