I Have AI Fatigue. Here's What I'm Doing to Overcome It

A cartoon man slumped over. Above him is a thought bubble with a drained battery.
(Image credit: Image by Mohamed Hassan from Pixabay)

I don’t think I’m alone when I say that I’m tired of reading AI papers.

I’m an adjunct college writing professor and teach online, and the prevalence of AI papers I suspect are AI-generated has grown to such an extent over the past year that it has changed my relationship with teaching. Although I’ve learned many great methods for detecting AI papers, searching work for signs of artificial intelligence is different than reading a student’s paper and connecting with it.

Teaching writing to inexperienced writers has always been difficult and, at times, frustrating. But there’s something deeply rewarding about taking the time to respond to a fellow human’s work, something worthwhile about trying to help them develop tools to better express their thoughts and feelings to grow as living, breathing, thinking, human beings. AI chatbots change this equation.

Now, instead of spending time trying to help students communicate, I’m looking for marks of the machine. When I see suspected AI use, I believe it is important to talk to students, but this is not only time-consuming it creates a situation in which I’m frequently conducting mini inquiries into student ethics rather than, you know, teaching.

Because of all this, for the first time in the 10 years that I’ve been teaching writing, I seriously debated taking off this semester from working with undergraduate students. Ultimately, I have decided not to do that.

Instead, I’m trying to change my mindset around AI work to regain a little bit of the fun and passion that I think the robots have stolen from teaching. Here’s how.

Overcoming AI Fatigue: Focusing On Quality Rather Than Authorship

One of the tells of AI writing is just a weird lack of quality. For a recent article I wrote on detecting AI, I was reminded that sometimes it doesn’t really matter if the paper is AI-generated or not. Poor-quality writing is poor quality.

As a writing instructor, my job is to point out how that can be improved. I don’t love the idea of critiquing robot work but there’s still a human at the wheel of one of these submissions.

So this semester I am embracing the idea that the human pilot of AI-generated work can still benefit from seeing the flaws in that effort. To do this, I’m letting students know that I’m grading AI-style writing harsher, so this will hopefully de-incentivize AI use overall.

Also, I will be able to spend less time assessing whether work is truly AI-generated or not -- if it is AI-style writing it's getting a lower grade. Overall, although it pains me to say it, if students are going to use AI in place of writing themselves going forward, there is value in helping them learn to get better results from their AI prompts.

Accepting That There Will Be AI Papers Submitted That Will Get Past Me

I think I’ve spent too much time focused on AI papers this past year. I’m still on the lookout, of course, but I’m also accepting the fact that I’m going to miss some AI work.

One of the schools that I teach at doesn’t allow use of AI detectors because of the well-documented accuracy problems, and without using one, there’s only so much I can do.

At the risk of sounding cliché, my new motto is, “Give me the strength to detect the AI papers I can and accept the AI papers I can’t detect.” Because of my new focus on the quality of the work rather than how it was created, it should matter less if an AI-generated paper or two slips by, as the students most likely won’t be rewarded for that.

Obviously, this feels like a surrender in some ways, but I’m reminding myself that despite my best efforts there were always students who I didn’t reach. That didn’t bother me if I was confident that I had done all I could reasonably do to help them succeed. I’m trying to get over my bias against AI work and extend that mindset to AI writing as well.

Reframing The Purpose of Writing

For a long time, I think English/writing educators have been selling the wrong message when it comes to writing.

A professor I worked for as a teaching assistant used to tell students that they needed to pay attention in their composition class because they’d need to write emails as part of their jobs. That rationale never made sense to me and makes even less sense now. First, most students in that class could already write coherent emails. Second, that type of work-a-day minutiae writing can now be adequately, and some would argue more than adequately, handled by ChatGPT or other writing AI tools.

That thinking was never compelling enough for students to cancel weekend plans to wrestle with the written word. Consequently, my new message to students is: You shouldn’t learn to write so your resume will look better or your cover letter will shine. You should learn to write because writing is one of the fundamental ways in which humans are able to express their thoughts, hopes, dreams, and more. Writing is linked to cognition and has changed the way humans have thought since ancient times. When we learn to write, the blank page can be a friend that we bounce ideas off of or a mirror of the deepest portions of our soul. Feel free to let AI write your cover letter, another AI tool is probably reading it anyway, but don’t let AI steal the fundamental human endeavor of writing from you.

I know that all students won't respond to this message but I'm dedicated to sharing it in the hopes that some will.

Erik Ofgang

Erik Ofgang is a Tech & Learning contributor. A journalist, author and educator, his work has appeared in The New York Times, the Washington Post, the Smithsonian, The Atlantic, and Associated Press. He currently teaches at Western Connecticut State University’s MFA program. While a staff writer at Connecticut Magazine he won a Society of Professional Journalism Award for his education reporting. He is interested in how humans learn and how technology can make that more effective. 

TOPICS