The impact of GenAI use on writing development

In a break from the many accreditation jobs I’ve been working through in recent months, I watched a video of David Perell interviewing ChatGPT’s Sam Altman. Perell’s YouTube channel is dedicated to adult writing of many varieties, and Altman is the CEO of ChatGPT and one of the biggest voices in Generative Artificial Intelligence (GenAI) on the planet. As GenAIs like ChatGPT continue to transform how we write and think in the education world, and how our students demonstrate understanding through writing, how Altman thinks AI will impact writing is relevant for every teacher of writing. You can view the full interview here (it’s very interesting), and in this post, I’ll share some thoughts about GenAI as a person who’s deeply interested in how young people develop writing skills throughout their schooling and beyond.

Altman begins the conversation by describing writing as an extremely important activity because it enables thinking and makes it concrete in a way that speaking cannot. Altman says, “If ChatGPT can allow people to do a writing-like activity and get better quality thinking out of it, that’s wonderful.” This optimistic take on the relationship between people, writing, thinking and GenAI emphasises a key potential affordance of AI for writers: Want to think more deeply and in new ways? Write with GenAI! Sounds pretty appealing, right? 

As a positive example, Altman says he recently witnessed a student using ChatGPT to help with their homework. As they worked, the student (apparently) thought, I’m kind of stuck, let me get unblocked and let me generate a bunch of other ideas. In Altman’s words, “The thing that came out was, I think, much better than anything someone would have come up with on their own”. In this example, the possible ideas generated by ChatGPT stepped in to help the student when they were stuck. It prevented them from needing to enter into the struggle of generating their own ideas: a struggle Altman seems to suggest here should be avoided. While it’s possible that “the thing that came out” was in fact better than what this student might have written independently, I’m not sure that this automatically makes the use of GenAI in this scenario a good thing.

Imagine wanting to run a marathon. Instead of training to build up your fitness and stamina, you ask a robot to run the marathon for you. Sounds much easier, and you could even win, but what do you lose in this arrangement? Sure, you might get a medal either way, and the robot might run the race faster for you, but without the struggle, there’s no actual development for you. 

A less optimistic take on GenAI and writing

Altman’s positive take on GenAI and writing is at odds with an impressively brief and thought-provoking alternative shared by scientist and writer Paul Graham in late 2024. For Graham, the impact of GenAI on writing will be profoundly negative, leading to a situation where, in just two decades or so, there won’t be many people who can write. He puts this down to the fundamental difficulty of writing (driven by the fundamental difficulty of thinking clearly). “To write well you have to think clearly, and thinking clearly is hard” (Graham, para. 3).” 

Writing well has always been difficult, but the new “escape valve” provided by GenAI means people no longer need to worry about entering into the struggle of writing (and thinking). While every writer you’ve met has likely already experienced the incredible efficiency of GenAI-assisted writing, those who have integrated it deeply into their workflows may not yet have thought deeply about the potential tradeoffs involved. As Graham (2024) argues: 

“The result will be a world divided into writes and write-nots. There will still be some people who can write. Some of us like it. But the middle ground between those who are good at writing and those who can’t write at all will disappear. Instead of good writers, ok writers, and people who can’t write, there will just be good writers and people who can’t write” (para. 8).

In a world where many people can’t write, collective human thought would be the major casualty. 

During Sam Altman’s interview, I was fascinated to hear him admit to witnessing a much less positive example of a second student completing their homework with ChatGPT:

“One of them basically just put in their thing and (it) wrote the whole essay, and I was like, appalled, because I knew that that was a theoretical thing that people were doing. … To watch someone do that and then get an essay that was, like, bad but passable out of it was, like, a real what have we done moment. It was visceral, in a way, like, I just hadn’t seen someone do it before.” 

Teachers are well aware of the risk of students simply using what GenAI spits out as their own work. While there are obvious issues around whose work it is and whether a student like this second one mentioned by Altman actually learned the content, the deeper issue I can see affecting every person who uses GenAI for writing is in its potential to stifle (or downright end) our own writing and thinking development. The struggle of writing that GenAI helps us avoid is a key aspect of learning. Offloading that struggle to a mysterious technological entity is, admittedly, a quick way to get something written, but if we aren’t careful, we may become entirely reliant on it for most of our future writing and thinking.

Here is a final quote from Altman:

“I was reflecting a lot on that (i.e., seeing the two students complete their homework with ChatGPT) and the first question was, like, a bad question, like, if you can just put something in and get a super interesting, or super passable, response, I think we’re just asking people to do the wrong thing. But if it’s something that gets them to want to think about a question differently and use the tool to get somewhere they wouldn’t have got on their own, that’s super interesting.”

We can all make predictions about how this will turn out. Clearly, even the guy running the biggest AI company has a variety of hopes and doubts! 

How to protect our writing and thinking skills

From my perspective, the wisest thing to do (no matter who you are) is to continue practising to write without the aid of GenAIs. If you use GenAI regularly at work, it might mean setting aside some regular time for your own writing (and maybe do some quality reading too if you’re not already). You could start writing a journal of your thoughts, get some ideas down for short stories, play with poetry, or write summaries of your main learning from different events. Like going to the gym or practising for that marathon, this kind of struggle will pay off and it’s what life is all about.

For me, a key outlet for my own GenAI-free writing is this blog (other than the odd AI-generated image). As you can probably tell from how it’s written, this work does not involve ChatGPT or any other GenAI. Does the writing kind of stink, Sure. And could I write many more posts by offloading the struggle to AI? Right again, but it wouldn’t be the same for you or me. It’s worth thinking about the GenAI-free outlets you can create in your own life.

For those who teach young people to write, I fear we can no longer justify the struggle of learning to write with the argument that, sorry kids, writing is crucial for your future work and there’s no magic get-out-of-jail-free card. There are many advantages of writing (and thinking) that are not easily accomplished in other ways. And there’s a lot at stake for students if they trade thinking and the struggle that generates and enhances it for quick and easy outcomes.

With so many students already finding it challenging to develop writing skills, I don’t think there’s any logical reason to introduce children to GenAI in the primary/elementary years. For those teaching secondary students, if GenAI is already part of your students’ learning, it’s important to consider and communicate with them the significant potential cost involved if they rely on it too much and too often. This is not just about plagiarism and who has done the work; it’s about the development of thinking and learning that is so wonderfully fostered through good old-fashioned writing. Call me a laggard, but the efficiency gains are simply not worth the tradeoff.