“Give a man a book, and you entertain him for a night. Teach a man to write, and you give him crippling self-doubt for life.” — Andy Weir, 2015

I recently completed a research paper for publication. It was the first time I’d written that kind of paper since I started using AI in my writing process. I assembled all of the books and papers I was going to use into a NotebookLM and got to work. It was a surprisingly painful experience but one that may have resulted in a deeper argument overall (only time will tell).

However, is there something deeper going on here with my thinking? I have long treated AI as a wonderful boon in making me time by removing a lot of mundane tasks from my plate. From creating materials for my classes on the fly to dealing with everyday correspondence, I could vastly shorten the amount of time I used to devote to such tasks.

What I didn’t realize was just how much AI would change my perspective on writing and where my ideas fit into the universe of ideas out there. I was writing a paper on the shift between a textual and post-textual society, but it turned out that I was experiencing some aspects of the shift while I was writing the paper itself.

As I describe in Discovering Digital Humanity, I learned to write digitally. What that means is that I never had to use a typewriter for serious writing. I never had to be systemic in my prep work. I got what I wanted to say onto digital paper. Then I worked backwards to edit it and source it. If I found gaps or contradictions, I corrected them as part of the editing process.

I learned to avoid outlining where I’m going (unless I’m writing something bigger, like a book) so that I don’t straightjacket my thought pathways. I just let my ideas flow and take me where they will, knowing I can clean it up in “post.”

For me, research is getting my brain where it needs to be to start and, as I described, filling in the blanks and correcting my thoughts as I go along. The most frustrating part of this process is working backwards through sources and finding new ones as I realize that I need to explore a particular aspect of my paper as I go along.

I thought AI would speed up this part of the task and it did. However, it was very good at opening new doors for me. That was a trap. Like Alice, I kept going through the looking glass because something interesting beckoned from the other side.

Instead of spending hours flipping through papers looking for highlights or flipping through books looking for notes, I would just search for them in my AI brain. Instead of spending agonizing hours struggling with the question of “what’s this paper really about?” I could use my AI brain to help me sort through the possibilities at speed.

That second process worked, at least in part. I gave NotebookLM the general premise about what I was going to write about, including data on the publication I was writing for. I also gave it my own writing over the last few years and then asked it to review my sources to give me a list of suggestions of how to frame the paper.

For instance, in the paper I theorized that ideas were bound into linear text and that visualization coupled with AI tools could provide new ways of looking at them. As part of my research, I came across the idea of “deep mapping” to contextualize information using GIS (Geographic Information System) tools.

How and why deep mapping is so effective would have bolstered my entire argument, but it would have added considerable length to the paper. That didn’t stop me from wasting a lot of time considering how I could add it. Meanwhile, the clock was ticking to finish the project.

I did not realize how much the “discipline” of my writing was being limited by my analog ability to explore sources to find complementary ideas. As we have all done, I always find things after a paper has been published that alters what I “would have” said, had I to do it over. This is a natural part of the learning process.

Using the AI, however, opened those “would have” doors before I finished writing. I found myself repeatedly exploring new ideas and adapting what I was writing on the fly. Then, I had to go back and fix the overall narrative to make it make sense. This had the paradoxical effect of lengthening the writing process.

AI has massively sped up this task but, at the same time, has opened new avenues of exploration. Authors know that what you leave out of your work is at least as important as what you put into it. I set out to write a 2000-word piece for publication. The last version topped out at almost two times as many.

As a writer, I frequently like to quote the Andy Weir quote with which I started this blog (note that I didn’t think of doing that until I wrote those words). We are constantly asking ourselves the question of whether anything we write is any good. The purpose of writing is to put yourself out there to everyone else to be judged. Otherwise, you might as well just journal your thoughts to yourself.

But most of us have a lousy sense of whether anything we write is good or not. Even successful authors like Weir constantly question whether they are creating things up to their own quality standards.

What I am trying to say here, is that I don’t know whether this process has yielded work that is better than what I wrote pre-AI. It has shifted the pain, not eliminated it. I like to argue that AI frees us from the mundane so that we can focus on higher level tasks that we never had time for before.

One thing this process has done is to give me a lot more to think about, both from the perspective of the writing process itself as well as the process I was writing about. Ironically, this topic morphed into a discussion of systems theory and the human process of adapting to new forms of sharing our ideas. I was therefore writing about what I was doing and doing what I was writing about. Talk about a rabbit hole…

Writing this paper has taught me that, even as a partner, AI reshapes how we think. Digital writing shaped how I thought about structuring my ideas. AI will do the same thing if I let it (and I need to). Ironically, it’s my experience as a writer that will make this transition harder.

AI opens doors for us to explore topics in a deeper human way. However, it also preys on our natural curiosity to explore. As writers and thinkers, we must adapt to this new reality. We can’t hide from it.

We all must learn to accept the incompleteness of our ideas. We can no longer entertain ourselves with the notion that our thinking will change the world and settle for just making it better before handing it off to the next thinkers in the chain. In this way, we can take pride in our role in building the human knowledge project without succumbing to the hubris of the “perfect” idea.