As I write this, I’m attending a virtual conference organized by ACES: the Society for Editing. One of the most popular topics under discussion is using artificial intelligence or large language models to generate text. Not surprisingly, ChatGPT comes up a lot (amazing to think that’s been widely available less than a year!). Also not surprisingly, editors have mixed feelings about a tool that so quickly and effortlessly generates convincing-looking text with few, if any, grammar errors.

Editors are curious, intrigued, cautious, and dismayed. That’s not just because of concern LLMs could replace human writers and editors. (Most editors I’ve chatted with don’t think that’s likely to happen. Yet.) But using tools like ChatGPT does raise a ton of questions and challenges about authorship; the ethics of writing and respecting one’s readership; efficiency and productivity; and the relationship between writers and editors.

My own stance as an editor, especially since I focus on academic writing, is this: I’m curious to know more about these tools—to test them out and understand their limitations, both practical and ethical. For example, Grammarly uses AI to recognize and predict textual patterns. It can help clean up rough drafts. But writers need to know it’s not perfect; for one thing, the tool can generate false positives that often, only an editor familiar with tone, nuance, and context can spot.

ChatGPT generates useful summaries and outlines that can serve as starting points for original work and can also pretty accurately list 10 cool places to visit if you’re in Brussels. But ask ChatGPT to write a short story or cite some sources, and you risk getting a bucketful of clichés or a series of “hallucinations.” (Yes, ChatGPT literally made up the names of several sources during one of my experiments, even though all were searchable!)

That’s why I’m cautious, as any writer or editor should be, about using these tools. I’m intrigued, but I don’t entirely trust them, and neither should you. Again, consider submitting AI-generated sources without checking them. That’s not only a breach of ethics but could have nasty effects on accreditation and reputation. Maybe I’m biased (OK, I am biased). But for me, all of this shows editors are needed more than ever.

Are you an editor or writer (or both)? What’s been your experience with using AI tools?

P.S. I used WordPress’s AI Assistant to get feedback on this piece before publishing it. The AI said I should encourage readers to share their experiences. Good point: so I added the question above. When I asked for more feedback, it said I should encourage readers to share their experiences. So, yes: please do!