I was asked several times last week if AI is going to affect my proof-editing business. So, I decided to get my head out of the sand and do some research. This involved signing up for the free version of a paraphrasing/editing/writing AI tool (probably not the first one that comes to mind). Disclaimer: Please bear in mind that I am not a researcher or scientist, so my research is crude and ad hoc.
The simplest way to conduct my research (as a non-researcher) felt like the most intuitive and obvious: to use the AI tool as if I were one of my clients, with my editing knowledge as the control and the most common grammatical issues that arise in my work as the research questions. So, using the grammar checker part of the tool, I started by typing in the following (grammatically incorrect) sentence: The dog ate the ice cream which was very cold. And it was corrected to: The dog ate the ice cream, which was very cold. I then tried a comma splice: The girl played with the kitten, then she went to bed. Which was changed to: The girl played with the kitten, then went to bed. So far, so good. And I carried on in this way, entering things that were, on a very basic level, grammatically wrong. And it did its job.
But, determined to catch it out and demonstrate that human editors cannot (as yet) be replaced, I entered something that was both grammatically and factually incorrect: I had a list of vegetables to buy; apples, kittens, bananas, hammers. It correctly changed the semi-colon to a colon (to introduce a list), BUT of course it was unable to detect the most basic problem of all with this sentence: that kittens and hammers are not vegetables!
And, just to be really sure, I typed in: Paris is in England All it was able to do was add the missing full point – Paris is in England. You see the problem here? These are obviously very crude examples (we all know Paris is in France), but so much of my work involves spotting subtle inconsistencies and contextual errors, as well as basic typos and grammar issues, that I would not at this stage be able to recommend AI for comprehensive, intuitive, intelligent, nuanced editing and proofreading.
You might be thinking: Well, of course she would say that. Translators, editors, proofreaders, writers and content-creators everywhere are understandably worried by what AI might mean for their work. But, having conducted this crudest of initial experiments (I will do more – better – research in due course), I honestly feel reassured (I was getting worried).
And beyond the technical issues outlined here, there is surely the simple fact that humans still want to deal with humans (don’t they?). Human contact is comforting, especially for those of us who work at our desks, alone, much of the time. When I need to sort out an issue with my insurance, for example, or book an eye appointment, I always prefer to speak to someone, a human, rather than a bot. I do think these AI tools are amazing (and I will use them from time to time to check my own knowledge, to enhance my work), but I now feel more confident telling people that businesses like mine are not under immediate threat. We need to embrace AI, and the incredible opportunities it offers, and to work with it where possible. We can’t stop it now – nor should we – but maybe we can try to use it to our advantage, to make our work, as humans, better and more accountable. Teamwork, init.