I do not use AI in the process of editing.
I’ve been asked about this more than once in the past month or two, and I thought it would be worth my readers’ time to explain why. There are three main reasons.
First: Language, ideas, and human expression are subtle and dynamic. In my experience, machines are incapable of editing with the sensitivity, creativity, and understanding that human beings can bring to the job.
This is important because AI’s accuracy, from an editorial perspectice, is spotty. Why is it spotty? Partly because AI models have been trained predominantly on unedited and unproofread material, but it’s also due in part to the fact that while languages have rules (with which a machine can be programmed) they also have a remarkable tendency not to follow them terribly closely.
Languages are living and evolving things, continually recreated by human beings in the process of their use. This is why human beings are best at knowing whether language is being used well, or even properly.
Second: AI has a pronounced tendency to flatten tone, phrasing, prose rhythms, and obliterate the idosyncrasies of usage and syntax that create the elusive phenomena known as “voice” and “style.” Insofar as human beings write — and read— in part because it is a way that we communicate with other human beings, our inimitable and unique voices are an irreplaceable core component that makes writing worth reading.
Third: I am extremely protective of my clients’ intellectual property. AI companies have been proven, repeatedly and in courts of law, to engage in the theft of intellectual property including that of copyrighted works. (If you’d like some background on this issue, here are some overviews.) In refusing to use AI as an editor, I am refusing to expose my clients to the risk that their unpublished work may be retained in part or in whole by AI and/or used for training AI models.
If you are considering hiring an editor, you may wish to consider these issues with regard to whether your editor uses AI while working on your manuscript.
As a private individual with an interest in the welfare of people and the planet, I have other reasons for not using AI. If you are curious about the state of play with regard to the known and emergent risks of AI, I heartily recommend the Massachusetts Institute of Technology’s MIT AI Risk Initiative website.
— Hanne Blank Boyd
p.s. Free subscriptions to Developmental Edits are always free. If you’re not ready to for a paid subscription, why not leave something in the Developmental Edits tip jar.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.