Summary

The author expresses a love-hate relationship with Grammarly, appreciating its assistance but finding its AI suggestions often miss the mark, especially in understanding humor and context.

Abstract

The article "Grrrammarly Can Be Grrreat for Comic Relief" by an author with a background in editing humorously critiques the AI-driven writing assistant Grammarly. While acknowledging its utility for catching spelling and grammatical errors, the author points out that Grammarly's suggestions are not always applicable, particularly when it comes to the nuances of humor and creative writing. The piece highlights the frustration of dealing with repetitive and irrelevant AI suggestions, the pressure felt by writers to conform to AI recommendations, and the potential for AI to misunderstand and misinterpret a writer's tone and intent. The author also ponders the implications of AI feedback on writing quality and the future role of human editors, suggesting that the AI's preference for uniformity may stifle individual writing styles.

Opinions

The author values human editors over AI assistance, emphasizing the importance of human judgment in editing.
Grammarly's AI is seen as lacking the ability to fully grasp context and the subtleties of humor, leading to inappropriate suggestions.
There is skepticism about the feedback loop involving Grammarly's AI and its user base, questioning whether the AI's learning process is truly beneficial or if it's led by the less competent.
The author is critical of Grammarly's tendency to favor consistency and uniformity in writing, potentially at the expense of creative expression.
The piece suggests a possible conspiracy between Grammarly and dictionary publishers, hinting at a commercial agenda behind the AI's suggestions.
The author expresses frustration with the persistence of ignored suggestions, likening them to vampires that rise again each editing session.
There is a concern that writers may over-rely on Grammarly, mistaking its suggestions for absolute rules rather than guidelines.
The author playfully proposes an "Editor Fairy Plan" as a sarcastic alternative to Grammarly's services, highlighting the perceived inadequacies of the AI.
The article questions the effectiveness of Grammarly's tone detection and its usefulness for writers, especially when the AI misinterprets the writer's intended tone.
The author implies that Grammarly's free suggestions serve as advertisements for its premium service, suggesting that the company could improve its marketing strategy.

I DON’T ❤️ RED LINES

Grrrammarly Can Be Grrreat for Comic Relief

But its AI and I don’t see eye to eye

Tony the Tiger image by Whoisjohngalt, CC BY-SA 4.0, via Wikimedia Commons

Expinions oppressed herein are the dual responsibility of my imagination and my typing fingers. Whatever resemblance they bear to reality can and shall be used against you.

That AI Works for The Man

Do you use Grammarly?

I did not, had barely heard of it, till I stopped working and started funning last year. As a past & present editor fairy, I’m a fan of human editors. Amazingly, my OTJ fairy edits were taken seriously. In the olden times, especially pre-internet, human editors were paid!

You’re no doubt aware our dismissals of “helpful” suggestions get back to Corporate. According to Grammarly’s blog:

AI systems also need feedback from humans. When lots of users hit “ignore” on a particular suggestion … Grammarly’s computational linguists and researchers make adjustments to the algorithms behind that suggestion to make it more accurate and helpful. *bolding mine*

How many users constitute “lots,” and who are they to judge? Are the inept leading the imprudent? Writers of humor may dismiss suggestions that nonhumorous writers take seriously, and vice versa.

Grammarly tries to help my title by suggesting a subtitle change to “but it is AI and I don’t see eye to eye.” Nonsense. Bad nonsense. ¹

I find red lines distracting, leading to NoIDontWant*THE*Here dismissal of the same inapplicable suggestions ad infinitum in editing sessions. Until I began wondering how that affects the AI’s misunderstanding of context.

The pressure I feel like a news editor

That’s how my sentence in the screenshot above would read if Grrrammarly ruled. It explains editing I’ve failed to notice online a lot lately — news editing must be so high-stress, only AIs will do it, missing a beat on punctuation to boot.

Oh, the pressure! I feel like a news editor. — Editor Fairy’s counter to Grrrammarly’s suggestion, except that’s not what I meant.

And, of course, Grrrammarly is cheaper than paying human editors. I bet all its employees get Grammarly Premium as a perk.³.¹

In this here story, a made-up word led down a pesky wabbit hole but Grrrammarly considers “this here” hunky-dory, along with hunky-dory & wabbit. I wouldn’t write them in a serious context, but that’s just me. Maybe they’ll imbue my story with a friendlier tone. Yep.²

IGNOREance would be bliss

This screenshot collage depicts my frustration. As I drafted the paragraph below, Grammarly suggested (1) **correcting** my made-up word “IGNORE-ance” to “ignorance”; (2) **adding an article** to its own canned language “Add to dictionary” so it would read “Add to the dictionary”; & (3) **correcting** “make up” to “makeup” in the context of invented words. As a capper, it **autocorrected** “make” without approval as I begged to “make it stop” while adding extra letters to MAKE and STOP.

While attempting humor, I make up and misspell words on purpose. What does my IGNOREance of them tell the algorithm? Will Grrrammarly add IGNOREance to its vocabulary of acceptable words if I repeatedly Dismiss or Add to dictionary? Then add a new sentence? OMG noooooo make³ it stoppp

The poor, put-upon Artificial Intelligence seems ill-prepared to deal with my issues. I’d like to know what Grrrammarly suggests to erudite writers. Does it come down to voice, and the AI doesn’t like mine? Grrrammarly must favor consistency and uniformity but sometimes I feel picked on.

And what in Horus’s name does Grammarly suggest to Hogan Torah? I’m afraid to ask him but am guessing:

GRRR banned him after the first-gen AI killed itself
he banned GRRR because he loves to interact with human editors

Editor Fairy conspiracy theories

For years, my job entailed editing, proofreading, and quality assurance. I cringe to think someone grooving to Grrrammarly may assume it’s singing the gospel, and their bogus input may “guide” me.

Even more alarming might be a conspiracy between Grrrammarly Corporate and the publishers of dictionaries.

Grrrammarly helps with spelling faux pas and burps if I get comma-happy. Interpretations of my tone are weirdly off base. Dare I correct them? Does an AI know from humor? Will ol’ AI understand that my “mistake” dismissal equals Erudite Writer’s serious acceptance of the same suggestion?

Asterisks rule; red lines drool

Asterisk blizzards, created by human editors, are easily shooed away forever by writers.⁴ Red lines added by context-oblivious algorithms arise from dismissal in each new editing session like vampires rejoicing in nightfall.

Why are we volunteering to teach Grammarly how to help us? I could help its AI more than it helps me. Until I come to my senses, my Primo Editor Fairy Plan is available for $42 per month, payable in advance.

If free suggestions are advertisements for those available with a paid subscription, I suggest you hire a new Marketing team, Grrrammarly.

¹ Screenshots herein are by author. This is the first of three GRRR stories. Here’s the second one:

Computational Linguists Are Cunning Linguists

Who else is thinking Grammarly’s Kool-Aid tastes bad?

medium.com

² Grrrammarly’s evaluation of my tone Monday vs. Tuesday. I began ranting, then regained my humor balance. I don’t care what GRRR thinks.

GRRR found me Worried & Formal on Monday and Informal, Friendly, & Confident on Tuesday. Also shown is my “completely Friendly” award — no asking “Which tones did we get right?” from Wednesday, which didn’t last long.

OMG, it’s Wednesday and I am now toned as completely Friendly. Grrrammarly is not my friend. Did those computational linguists seize control?

Thursday, I ditched unwanted friendliness while gaining nerves & disapproval. Baby steps. What tone would help Grammarly improve its current tone suckiness? Not that I care.

Thursday’s tone is Sad, Nervous, & Disapproving. As if I cared.

Friday update: GRRR gave up on me; no tone, only advertising. Saturday, when I split my story into three, we returned to Worried & Formal here.

“There are 20 advanced suggestions available only for premium users.” I reached 21 before my GRRR frustrations morphed into a triptych.

³ Note how Grammarly surreptitiously changed my word while I was on a roll. This explains “I could swear I typed X! Why is it Y?”

³.¹ Happened again when I typed perquisite on Friday. GRRR changed it to prerequisite behind my back. I gave up and wrote perk. Now we know GRRR is on board with dumbing down. Dumbing down is of the devil, IMO. “They” could’ve at least spelled it perq.

Do unto editors as you would have them do unto you

⁴ If interacting with note-leaving (NL) editors, please don’t dismiss notes until your darling is published or removed from the pub unless they say it’s OK.

How do you feel when your darling is published by an editor who didn’t leave notes, then you realize they edited something? Without rereading, it’s hard to figure out what, if anything, changed.

That’s how NL editors feel when they check back on your note-free darling, multiplied by the # of babies they’re on point for publishing that day. Your darling’s publication is delayed when we can’t simply scan for lingering birth defects. I am so sorry. Them’s the breaking waves.

Thank you, writing buddy Smillew Rahcuef, for being way more helpful than Grammarly.