ARTificial Intelligence | Descriptivism and A.I. Writing Assistants
Like it or not, machine learning models—or A.I.—are quickly seeping into every technological aspect of our lives. And along with that ubiquitousness comes a wave of hype and panic; if it isn’t being positioned to achieve sentience and take over the planet, it’s in headlines everywhere as replacing human workers at every level of industry—and in creative fields in particular. But beyond the hype—the maybes and what-ifs and possible far futures—the fact remains that machine learning models are being used in a growing number of sectors; not least among them publishing.
This series dials back the layers of hype and fear to look not at how important A.I. may be in the far future, but how it’s being used right now in ways that directly impact the publishing industry. Inevitably, some of the talk about A.I. will end up being nothing more than hype, as we’ve seen from a cycle of tech bubbles in the 2020s. But in the meantime, its capabilities right now are impacting the publishing landscape in ways that are unlikely to go away anytime soon.
The presence of A.I. in editing software and word processing programs has become inescapable. Microsoft Word, the program I’m writing this in right now, uses machine learning to suggest fixes for grammar and spelling, and recently rolled out even more A.I.-driven features, including the integration of generative text models. It’s not the only one, either; Google Drive has increasingly been selling A.I. in its marketing as a tool at the disposal of writers, and cloud-based grammar correction software Grammarly has become a household name in recent years. Some of these functions aren’t what we conventionally think of as artificial intelligence. But they’re all doing similar things: looking at the words used in a given document or chunk of text and comparing them to large datasets to see how these words are usually used and what is most likely to follow them.
Descriptivism is not new to the grammar world, but a human writer or editor can shift between prescriptivist and descriptivist approaches on a case-by-case basis, determining what might be a shift in the usage of a word and what’s a common error that still isn’t accepted by professionals. An A.I. isn’t capable of making that distinction. This isn’t news; we’ve all gotten incorrect or clunky grammatical suggestions in a word processing program or had a proper noun auto-correct to a similar-looking word. Luckily, a seasoned editor can look at a grammar suggestion in a Word document and make their own decision about whether their phrase is prescriptively or descriptively correct. But not every user has that kind of experience. And while whether a sentence should read “me and Janie” or “Janie and I” may be something people can find with a quick Google search, as machine learning models become more ubiquitous in word processing software and broaden their functions to include full-on generative text and suggestions on style and tone, it’s worth thinking about whether we trust a computer program to make these kinds of distinctions for us.
Do we really want the current and future generations of writers using this technology to learn how to write—and how to edit—based on what an A.I. sees as the most common way of ending their sentence? Descriptivism has a place in grammar, sure, but can we apply those same principles to tone? Is there even a “correct” way to style a piece of writing, or just a more homogenous one? Editing software has become an important part of our lives, and I don’t think that it’s going to wipe away all traces of human creativity. But it is capable of re-training the way we write—just think of the last time that you let a typo stand because you knew spellcheck would catch it, or accepted all of Word’s suggested edits because you didn’t have the energy to read over the piece again before a tight deadline. Applying a sweeping descriptivist approach not only to grammar but to writing as a whole has the ability to reinforce the idea that how the majority writes is the better or more correct way, and those assumptions do matter. A descriptivist approach to grammar is populist; a descriptivist approach to writing may push down voices that don’t fit the conventional mold.