There’s a common phrase in many freshman composition classrooms that goes: “To write is to think.” The notion here is simple: The more you organize your thoughts through writing, the more refined, exact, and fully developed your thoughts become. George Orwell, the great British novelist and essayist, wrote to this effect in his great 1946 essay “Politics and the English Language.” Orwell points out how poor thinking leads to bad writing, which in turn leads to worse forms of thinking. And on and on it goes.

I’ve personally found this statement to prove true in my own work, and I know so many other writers will say the same thing. Oftentimes, we don’t even really know what we think until we are forced to write it out or speak it out loud. So, it came as some surprise when the New York Times published an op-ed by the chair of the philosophy department at Williams College, Joe Cruz, on how writing, contrary to common assumptions, is not actually vital to the process of becoming a better thinker. He even goes on to say that writing professors shouldn’t ban students from AI use in their papers. Cruz writes,

Writing remains an essential life skill, and we will continue to do it. A.I. will write summaries of our thoughts and may contribute some of its own. But we shouldn’t always force students to maintain a written record of knowledge. Universities have forgotten that archiving and thinking are separate.

Cruz uses oral exams to test his students’ knowledge of the course material and says that he’s been encouraged by the results. But is inviting AI into these types of environments really conducive to growing as thinkers and writers? Many would say no.

Journalist Matt Taibbi chimed in with his thoughts about the op-ed. A longtime critic of the AI creep in higher education and the writing world (although he certainly doesn’t reject AI wholesale), Taibbi writes,

The writing process does not work without the element of pain and rejection from a human audience. The whole idea is to put a part of yourself on paper and allow yourself to be vulnerable by showing your writing to another person. A negative reaction can be crushing, but it forces writers to confront the reality of how others view them. It might hurt the first time, or the first thousand times, but the human element is how you learn.

Taibbi is responding to the idea that we can just “bounce ideas” off of AI instead of running our work by another person for feedback. For Taibbi, and many others with a similar pushback, writing is not only a human activity but needs to be vetted and discussed among real people. The process can be gritty, but it’s worth it.

Cruz seems to think that focusing too much on writing can isolate students and keep them from exchanging their ideas in open dialogue. Taibbi reminds us that writing is actually a highly social and conversational activity. We write to communicate a message to an intended audience. So much rich conversation results from the provocative and insightful things we read. Yes, AI can write summaries and give us a sort of cognitive short cut. It doesn’t necessarily follow, though, that taking this route leaves more room for dialogue. For those who want to grow and develop as thinkers and communicators, writing is still a time-tested option.