Premise

We’ve all done it. Back in grade school (or maybe even in college too) we combed through our papers to try finding synonyms for vague, repetitive, or bland words. Our goal? Make our writing sound more impressive. After all, why sound just “good” when you can sound “vociferously tautologically adept” instead?

And so while we’re at it, why also not just have the bots do the search and replace for us? Why scour through our own work, look up new definitions in an online thesaurus, and replace each one manually? There’s way too much doom-scrolling we need to get to. And now, you can just hit a button, and, voila! A fancier sounding document is born!

So what’s the problem?

Well, the answer is, simply, because we should have never been doing what we were doing the first place. Not because it’s necessarily “wrong” in any particular moral sense. But more because it’s simply not representative of ourselves, and what we’re truly trying to say. And worse, it doesn’t teach us how to build the cognitive connections in our minds to think about why we’re saying whatever we are. And that’s a skill that far transcends writing.

Examples

Failing to Improve Understanding

Take the following sentence as an example of a typical line from a student’s paper:

“The author is trying to say that it doesn’t know the idea as good as it should.”

If this line seems confusing to you, then you’re not alone. To the writer of this sentence, the meaning might be clear. But that’s because that writer knows the intention behind the sentence. We, as audiences, do not. What does “it” mean here, for example? Do both “it” words refer to the same identification (i.e., the author? an idea?).

Obviously, you might be tempted to just let AI figure out the answer. But that’s the problem; like we just said, only you know (and if not, you should probably try to figure it out) the full context of what you’re trying to express. AI might say it knows the answer, but it can be very, very wrong in cases just like this.

“The author is trying to say that he doesn’t understand the concept as well as would be ideal.”

If this second version seems much clearer to you, then you’re not alone! Obviously, as the writer here, we had to imagine clarifying some of these vaguer key terms: “it,” “know,” “idea,” “well,” and “should.” But consider if you outsourced this clarification to an AI; how would it know that you meant, for example, to specify that you were trying to distinguish “ideal” from “should” versus another synonym? The answer is that it doesn’t—and never will.

But now you know, and most importantly, you have thought about it! I.e., your brain (a muscle) is literally now cognitively stronger than before. And that you can use in all other aspects of your life (academic, personal, and professional).

Failing to Improve Tone

You may also be tempted to use AI to “improve” your tone—and this is probably the most common usage of AI that students implement in this regard. E.g., a teacher returns your paper to you and says something like “Great job! But maybe work to make your tone more formal.” Or “A more professional tone would make this piece really shine!”

Oh, okay. You know who can do that? Why, any AI bot, of course! Just pop your paper in, ask it, “Yo, can you make my tone sound better?” And BOOM! Tone enhanced.

So what’s the problem, you might ask? Well, in a purely objective sense, maybe not much in some cases. But again, consider our previous points about thinking through these considerations to improve your brain itself. And that’s not hyperbole; scientific research actually shows an increase in cognitive functioning through these types of processes. I.e., if you outsource these processes, you’re also outsourcing your brain’s very aptitude.

Just to give a quick example. Look at the following examples and note the differences in tone (let’s assume we’re trying to make the tone sound more “academic/professional”):

Original: The speaker’s point about why students using artificial intelligence is bad can definitely be connected to other arguments where scientists point out the problems with letting AI fix your writing for you.

AI Edited: The speaker’s argument about why students using artificial intelligence is detrimental can definitely be connected to other claims where scientists point out the problems with letting AI enhance your writing for you.

Student Edited: The speaker’s argument about why students using artificial intelligence is problematic can definitely be connected to research where scientists point out the issues with letting AI “fix” your writing for you.

So, obvious question: what’s the difference between the AI edited and student edited versions? Well, just look at the specific word choice changes. The AI one is definitely fancier. But the student one has more subtle changes that, in this case, reflect the emphasis with which the student wants to express what they actually feel about this commentary. E.g., quoting “fix” is a harsher tone of indictment (ironically) in opposition to letting AI edit your work for you.

Obviously, any of these versions are objectively “fine.” But only one reflects what and why you actually think what you do. And that’s the key: to think.

Further Considerations

• Sure, AI can make your writing sound “fancy.” But is that really all that you want? Worse, sometimes AI mistranslates what you’re trying to communicate—just like in the aforementioned examples.
• Use AI to make suggestions. This is obviously a potentially tricky and slippery slope. But while AI might be able to point out ideas that you hadn’t considered for further clarifications, only you can make the final call yourself.
• It’s probably fine for “proof-reading.” In essence, there’s often nothing wrong with having AI scan work for errors. I.e., we’ve had spellcheck for decades now—which is essentially AI. But again, there’s a stark difference between asking a program to scan for missing punctuation or erroneous spelling and asking it to write a work or make your work sound like someone who you are not.
• If you’re not sure, ask your instructor. This guidance obviously applies to students in classes where these are concerns are very much real. The last thing you want is to be accused of plagiarism, so when in doubt, ask ask ask about the proper and allowed usage of any and all AI tools for your course. Usually, as long as you do this, you should be fine!