When I hear the above from a prospective client, it’s pretty clear that our collaboration will not go forward.
It will not go forward because—experience would indicate that this is generally case—the client who asks such a thing doesn’t understand the craft of writing, and therefore does not respect it. Should I collaborate with him, he won’t respect my work, either. He won’t be able to, because he doesn’t understand what it entails.
What such a client is looking for is a quick result at rock-bottom price. He thinks that using AI will reduce both his work and mine. But this isn't true.
I usually explain that since it won't reduce the amount of work I’ll have to put into it, I certainly will not offer a reduced price to clients who use AI. On the contrary, since the work will have to be largely rewritten in order to be humanized, I should like to be paid as a ghostwriter rather than a proofreader—that is to say, a great deal more.
Sometimes, I’ve taken the time to explain further: AI produces writing that sounds like AI. To any literate person, it’s as recognizable and as cringeworthy as an AI image. Not only does it sound vague, cheesy, and fake, it uses catchwords, clichés and recognizable patterns that quickly become boring to any attentive reader. Just as AI images frequently feature humanoid figures with deformed limbs or six digits, or introduce anachronisms in “historical photos”, so AI text mixes up the logical order of sentences or leaves out key information. Sometimes its “sentences” lack a predicate. Sometimes it’s impossible to identify a subject.
Correcting such errors takes a great deal of time and skill from a copyeditor. When we run across errors produced by a human being, there is (usually) a method in the madness—we begin to realize that our author has a very elliptical style that sometimes leaves his meaning obscure, or that he has developed a bad habit of unnecessarily keeping the verb from his reader until the last possible moment. Or perhaps we recognize that he’s thinking in an Italian or Spanish language pattern. Once we understand what he’s doing, we can intelligently intervene with a word here, a punctuation mark there. We identify the author’s thought, and then our work consists in revealing it with sensitivity and fidelity. Our work is an act of respect for the author, for his thought, and for the craft we both practice--not to mention, an act of respect for the reader. It’s a human work, produced by and for human beings.
With AI, it is not so. There’s no method in the madness. No thought lies behind the machine’s mistakes. There is no “there” there. Querying the “author”—that is, the client who sent me this mess—yields no useful information. He doesn’t know what he meant—because he didn’t actually write it. He didn’t think.
And that means that any humanity in the piece, any thought, is mine. It means that I’ll be the uncredited author of a piece of work for which I’ll be paid nothing more than a small sum as a “proofreader”—while the ostensible “author” enjoys the fruits of my labors.
No, thanks.
Photo by Unseen Studio on Unsplash