By David Ebenbach
Given the new capabilities of generative AI tools, a lot of writers are asking where and when (if at all) they should be using these tools to help them in their writing. On the one hand, AI causes an awful lot of problems, but, on the other hand, we’re being told great things about what these algorithms can do. So, how do we approach this technology? How might AI help in your writing process, and how might it hurt that process?
1. Idea Stage
It’s very tempting to want to use AI tools at this early stage, where you’re looking for creative ideas. After all, the blank page is awfully daunting. And yet using AI here would be a mistake, and not just because of its significantly negative impact on the environment; AI tools are, by design, an aggregation and rehashing of old writing and ideas, whereas you are a being that has never before existed, and therefore have a unique perspective that’s crucial to creativity. Some people go so far as to define creativity as authenticity, so this is a poor choice of a moment to give the reins to a machine. Instead, find a way to be yourself. Do some free-writing. Take a walk. Read widely, so that you avoid sounding like someone else in particular, allowing yourself to be provoked into your own distinctive thoughts. Ask yourself: what matters to you? What are you chewing on? Bring yourself to the task.
2. Drafting Stage
Similarly, the drafting stage is an essential opportunity to allow your own ideas to develop and to produce new ideas by connecting to other things that matter to you. The initial ideas are just a doorway into what you have to say—the drafting stage allows you to go through that door to explore. And so it’s best to hold off on using AI here, too, and not just because it steals writers’ intellectual property without permission; drafting a new piece of writing is a journey of discovery—full of happy accidents, challenges, revelations, unexpected turns that can make your piece (and you) wiser and stronger if you don’t hand the process off to a bot. Get in there yourself and make things happen.
3. Revision Stage
I personally hate revision—it feels to me like breaking a vase into a thousand pieces and then trying to put the pieces back together to somehow make a better vase that is also somehow still watertight. I hate it so much. And so this ought to be a great place to bring in AI. Unfortunately, it’s not. Never mind the fact that generative AI is threatening the very foundations of truth as we know it; the real reason to avoid it during revision is that, as much as I hate it, revision is actually another learning opportunity. Revision is a chance to look at your raw piece and see all the different things it’s pointing at in its chaotic rough draft way, and then make choices about which one(s) you want to follow, so that you can really pursue something you care about. This is the stage where you find out what you’re writing about, and you can only do that if you’re in the weeds yourself.
But what about using AI to give you feedback on your piece? Some folks are doing this, I know, but I don’t recommend it. First of all, if you’ve read writing produced by generative AI, a lot of it is…bad…so I personally wouldn’t trust its suggestions. Also, AI is famously prone to telling us what we want to hear, so it may not give us the tough-love advice we often need as we hone our work. And, finally, this strategy means engaging with a machine rather than with other human beings. Why feed the algorithm when you could instead be building community?
4. Submission Stage
Well, at least submitting your work is semi-mechanical, right? Why not bring AI to bear here, helping you choose targets and write your cover letters and so on? Well, first of all, AI makes a lot of mistakes, so you’ll have to doublecheck everything it tells you, which might mean saving less time than you expect. But there’s another reason why you might want to do this work yourself. Making the submission stage mechanical is the same as making things transactional: grab a list, blast off, wait for results.
The alternative is to become a contributing part of the literary ecosystem by reading magazines and journals, whether online or via library copies or bookstores or paid subscriptions. In this way you’ll find venues that share your sensibility and goals, and you’ll serve the community as one of its most important elements: a reader. After all, you want your work to have an audience—so why not do the same for others, rather than having a bot scrape websites without any appreciation for all the creative work to be found there?
5. Miscellaneous Stage
After all the above is done, maybe there’s finally an opportunity for generative AI to come into the picture. You could ask it to, for example, create a cartoon dog congratulating you for all of your hard work. The writing is done, so ordering up a cartoon dog won’t allow you to skip any of the essential stages of your own creative process. Of course, using generative AI means supporting massively wealthy tech companies and giving them more control over our culture, as well as encouraging unemployment and the perpetuation of bias, so I tend to forgo the cartoon congratulations myself.
More earnestly, the real point here is that your voice is the voice that the world needs. It needs your ideas, your rough drafts, your careful reworking, and even your engagement with the publishing world. In other words, there isn’t one piece of the unbelievable gift of human creativity that is better off in the hands of AI.
Bio: David Ebenbach is the author of ten books of fiction, poetry, and non-fiction, winners of such awards as the Drue Heinz Literature Prize and the Juniper Prize, among others. He lives in Washington, DC, where he teaches at Georgetown University. You can find out more at davidebenbach.com.
