During the last several years, I’ve slowly been completing a copyediting certificate program. Part of the motivation for this additional training is that I realized editing is a skill that can benefit from training. A lot of academics think editing is just “proofreading,” by which they mean “catching typos” or fixing bibliographies. However, I’ve come to understand that copyediting involves much more than spotting the errant comma. (And, for what it’s worth, “proofreading” is different, as it involves checking proofs after the typesetting stage. Copyediting focuses on the manuscript in preparation.)

So, I’m now one course away from completing a certificate. The process has helped my own writing process immensely. In the past year, especially, I’ve developed a more systematic approach to writing, revising, and preparing manuscripts for review. And along the way, I’ve also learned that everyone makes mistakes, especially in their own work, and this is why having an additional set of eyes that isn’t close to the work can be so helpful.

When it isn’t possible to hire a professional, there are some ways to simulate an outside perspective on your own work, which is what I’ve been doing as I’ve been editing a four-volume book series. In addition to checklists, which I might post about later, I’ve found Word macros to be useful. (These won’t work for writing in LaTeX.) Paul Beverly’s website has helpful advice for getting started, so I won’t reinvent the wheel here. Instead, I want to consider:

Why would you want to use a Word macro in the age of AI?

I know a lot of academics who think that, given the rise of large language models and user-friendly interfaces like Claude or ChatGPT, human editing is a dying art. They can simply ask an AI to edit their manuscript for them. However, I think there are good reasons to use macros (and perhaps, to resist AI editing at key stages and involve human editors).

• Macros do not require you to upload your manuscript to a tech company. I was at an “AI and Editing” Zoom panel recently, and Jenni Fry, who manages the Manuscript Editing Department for the University of Chicago Press Books Division, kept emphasizing concerns about intellectual property. While there are ways to run local language models, I imagine most academics are just going online to ChatGPT and uploading their manuscripts. But this means risking that your intellectual property will be used for training data. There’s a lot of uncertainty about what tech companies can and can’t do with your texts, and there’s also, in my mind, doubt about the extent to which they follow even their own stated commitments.
• Macros are transparent, not a black box. The macro’s source code, Visual Basic for Applications (VBA) is not difficult to understand. That means it’s transparent how the macro is giving you the results you see. In comparison, LLMS run on hidden information, sandwiched between layers of complex neurons made up of many-dimensional vectors that are beyond the capacity of human beings to interpret and understand. While this makes them powerful, it also makes them opaque. Of course, the process of human judgment, picking one word over another for stylistic reasons, is also a kind of black box. But at least some stages of the editing process can be made visible with macros.
• Macros run on your own power source. The extent to which the rise of LLMs is already negatively impacting the environment (and how much it threatens resources in the future) is controversial. I am not sure where I stand on the issue, personally. It’s possible that individual editorial tasks are not the kinds of uses that, even cumulatively, contribute to environmental resource depletion. However, if this is a concern for you, macros, insofar as they run in MS Word, are not contributing to the resource demands of LLMs.
• Macros require you to think about editing. The final reason that I am integrating macros into my process is that they require me to think about what stage of editing I’m at, what aspects of the document I want to evaluate, and which changes I want to make globally or locally, bit-by-bit. It’s a small way to resist the frenetic pace of automation, but perhaps an important one. Automation is not a bad thing. I’m not a better editor if I change each instance of an incorrect spelling myself or have a computer do it for me. What matters is that I do not introduce new errors and that I remove existing errors. Since there are so many macros (see a list here!), you can’t, and shouldn’t, just run them willy-nilly. Instead, using them requires some thought about how automation would save you time. Also, if there are mistakes you often miss (or errors you’re prone to introducing), reflecting on these might lead you to a macro.

While an AI tool can check grammar, fix citations, even rewrite for clarity and give “developmental editing” reports, it cannot make the judgment for you, the author, about what the best choices are for your manuscript. What I like about using macros is that they can automate the process of analyzing and revising your manuscript so you can make informed, human choices.