I know, I know, editing is expensive! It’s intimidating! It’s vulnerable! It’s time-consuming! Trust me, I really, really get it. And when AI is right there, promising that it can polish your work, fix your phrasing, and catch any mistakes, it can be tempting to just use that instead.
I am not a fan of AI, as I’ve said elsewhere, primarily for ecological reasons and because the powers that be seem suspiciously eager to replace human ingenuity and artistic production with bots that they don’t have to pay. Also for data privacy reasons; I use Proton AI when I do use it at all, and even there I don’t share anything personal or confidential. Note that I never, ever use AI as part of my editing work.
But that said, I know that folks have questions about what AI can and can’t do, and I think it’s important to have a good grasp of how AI can help or hurt your writing. I have played around with it myself and I have also worked with clients who have used AI in their writing, to various degrees. (I know not all editors will work with clients who use AI, but my stance on this is that my work as an editor remains the same, however the text was produced. That said, I don’t work in fiction, which is a different kettle of fish)
What AI can do
Gen AI is generally—although not always—pretty good at collating and synthesizing large amounts of information. So, for example, when presented with a block of text or a large amount of data in some other format, I have found that it’s fairly good at going through it and picking out keywords or noticing patterns. It will sometimes miss critical information, or it will draw a conclusion that you yourself wouldn’t have drawn, but it’s not bad.
Note that I, obviously, would never recommend this type of process for an academic study. However, for your own writing process purposes, this means that AI can produce a fairly decent reverse outline, and can potentially spot holes in an argument, if the holes are relatively simple. I like reverse outlining myself, as I am someone who needs to write first, structure later, so doing a reverse outline (outlining a paper after you’ve written it) helps me spot gaps in my logic. If you feel comfortable inputting your work into an AI system, then this is a good tool (even if you do feel comfortable, I urge you not to use ChatGPT or Gemini for something like this. Use a secure system like Proton, or even Claude.).
What AI cannot do
AI is not a good copy editor. It can do the reverse outline, and it can perhaps spot some glaring gaps in logic. However, AI, even when prompted to fix spelling and grammatical errors, will miss some. It’s really astonishingly bad at catching mistakes; and even old-fashioned spellcheck is better than generative AI is.
I’ve run quite a few tests on my own, and I’ve copy edited work that clients have run through AI in an attempt to fix the language before the document gets to me, and even these are faaar from perfect. In my experience, good human feedback is still the gold standard.
There is also the danger of Gen AI rewriting sections unnecessarily. I have had repeated issues in my various experiments with Gen AI where it consistently adds sections or rewrites sections of the inputted text, even when I’ve explicitly told it not to. This is a concern if you’re trying use it for copy editing, since there could be sections added or conclusions made up that you don’t even realize, which could in turn jeopardize the quality of your work.
If you are going to use Gen AI tools, I think it’s important to know how to use them safely, effectively, and ethically. From my point of view as a professional editor I can see multiple ways in which it can be effectively used in scholarship; and have also witnessed the many ways in which it can be detrimental to both scholars and scholarship as a whole when used thoughtlessly.
I hope this is helpful. Good luck on your writing journey, and if I can help please get in touch or consider joining my free Discord community, the Academic Support Hub!