Recently, I began work on a presentation I will be giving soon.* And as I was writing something very strange began to show up. I was getting the little wavy red lines under my words. Words that were spelled correctly, and that in good standing with the standard grammatical usages.
When I clicked on them, I was getting a notice about conciseness. Meaning the system is judging my words against some arbitrary standards of narrowly defined language (e.g. “business English”should be written and applying it to all writing) and the dumbness of that idea aside, lets talk about why this “AI tool” should not be trusted.
Concision is Not Always The Best Path Forward In Writing
Not all writing should be concise. People use word processing and presentation making software to communicate in a variety of ways. Yes, you might be writing a report for work. But you might also be writing a short story, or making a presentation (which is, the ONLY thing you usually do with Power Point) you may be worried about more than concision. Like creating a connection to your audience or being clear about your points, even if that makes your text a little bit wordier.
It also, is a way of erasing identity. While I don’t have the colloquial phrases and beautiful circuity of a southerner, or the propriety of those who speak The Queen’s English, I value and appreciate these things. People are partly defined by these little quirks in how they interact with and describe the world, and standardizing and removing them will take something away from that beauty and uniqueness. And for what? To make things more acceptably corporate?
That’s not a worthy trade.
The AI is Poorly Trained, and Misses Out on The Nuances of Words, Mostly Picking a Shorter Synonym
The AI isn’t very I in this case. It seems to just take what it things is a shorter synonym of the word and suggest that you put it in. It does not worry that much about if the words have the same meaning, let alone the same precise meaning. Yes “fire” might be shorter than both “candle flame” and “a two block conflagration” but they don’t really mean the same thing at all. Words have specific meanings for a reason, we are trying to communicate something and boiling it down to the shortest option does not allow you to convey that.
The AI Can Make Sentences Grammatically Wrong and This Will Make Worse Writers of The Users Who Rely on It
As you’ll see in some of my examples below, the system can actually create errors in the grammar of the document. Now, if you are already a strong writer, then you can spot these incorrect corrections. But if you are younger, or an ESL writer, then you might believe these “corrections” and they will end up shaping what you think is correct English.
On the whole, we should be striving to help make better writers with tools like these, and any attempts that could result in incorrect assumptions should come with context.
I Want to Pick My Own Words
And finally, I just want the ability to choose my own words, without being judged for it. If I want to say “The dog is brown” or “That dog, who is brown” or “That dog happens to be brown” should be my choice. I know I’ve said this a few times now, but writing is an act of communication and in some ways one of the most enduring ways we have to express ourselves. I want the ability to make those choices for myself, without an automated system trying to babysit me.
Turns out, for those of you who also want to throw off the yoke, that you can turn this annoyance off. All you have to do is turn off all of the grammar checking completely. Because that’s a fair way to treat your end users!
*And might I say, bless the virtual conference! It makes it so much easier and accessible to do multiple speaking engagements in a year without the costs of travel. I hope these never go away post COVID.