The Funneling Down of Language
How gen AI impacts our language and what this means for writing teachers.
Recently, posted on LinkedIn about what I have been calling the funneling down of language.
At first I thought it was just me, [b]ut If you use gpt 4o a lot [y]ou will also notice this type of writing all over LinkedIn:
It’s not just X, it’s Y.
We didn’t X, we Y.
What do I mean by “funneling down”? The more that people use AI programs to write their ideas for them, the more that all the posts and podcasts start to sound the same. It’s like a funnel in that you start with myriad voices at the top of the writing process, but end with one of two or three options by the end.
Figure 1 gives us an idea of what this might look like. And the more often people publish this, the more people read it and then, even if they are doing their own writing, they may choose to mimic what they believe to be a successful writing style. So we end up with a whole bunch of weird sounding writing that seems to perpetuate its own growth.
I don’t love that.
And then there’s the post from noting that some of her students have started modifying their own writing so that they aren’t flagged by AI detection tools. Students are learning that if they use this word instead of that word in their own human-generated writing, they will be less likely to deal with an academic dishonesty case against them.
Sit with that for a minute. Our students are CHANGING THEIR VOICES because of AI-generated text. Not because they feel like these other words sound more like them, but because they feel like these words sound less like a bot.
And why wouldn’t they shift their language choices? We see lists like this one from at NI Schools Hub entitled “Top 20 Common AI Phrases and Words” or this joking on from with a ChatGPT Bingo card filled with words that people commonly associate with gen AI, which can lead people to believe that if a writer uses certain words or punctuation, the writing has to have been done by a gen AI program. But as noted last year in her small study, neurodiverse writers might also be flagged for their writing because it may sound bot-generated.
English composition and writing classes
So why does any of this matter? Who cares if we all end up sounding the same? Well, writing instructors care, and not just because we don’t want to read 25 papers using the same vapid and meaningless words. No, we care because we want students to learn about the vibrancy and beauty of their own languages.
Everyone has a home dialect — a set of words and inflections that they use to communicate with friends and family. Many of us also have professional dialects that we feel are more acceptable in the workplace. For example, at home I might drop the -ing off the end of words, or my Valley Girl (thank you 1980s) inflections might show up when I am really excited about something. But at work, those -ing endings are firmly in place and I almost never use my Valley Girl voice unless I am in a small group of work friends.
Our students come equipped with a few dialects as well. Some of these are regional or national differences such as how many people now drop the “to be” verb tense in phrases like “My car needs fixed.” Other differences come from people speaking a second language and doing their best to translate ideas across those language barriers.
And some differences may come from spoken language within families or from how K-12 English class instructors told them to write. Some of these voices may be more or less authentic to the students. Those less authentic voices (like when someone tells me that I need to sound more academic in my writing) are harder to communicate in, and our ideas may get lost in the shuffle of trying to find the right words for the situation. This is where gen AI tools might come in handy for helping to close a communication gap.
Maybe our students feel the pressure to get good grades; maybe they equate the sometimes vapid and meaningless prose that gen AI tools output with “good writing” because there are big words. This is where writing instructors can step in and fill the space that gen AI tools are trying to fill. We can remind students that their own authentic voices have power and that we genuinely want to hear what they have to say in their own words.
In 1974, the Conference on College Composition and Communication (CCCC) wrote stating that students voices matter and should be respected:
We affirm the students’ right to their own patterns and varieties of language — the dialects of their nurture or whatever dialects in which they find their own identity and style. Language scholars long ago denied that the myth of a standard American dialect has any validity. The claim that any one dialect is unacceptable amounts to an attempt of one social group to exert its dominance over another. Such a claim leads to false advice for speakers and writers, and immoral advice for humans. A nation proud of its diverse heritage and its cultural and racial variety will preserve its heritage of dialects. We affirm strongly that teachers must have the experiences and training that will enable them to respect diversity and uphold the right of students to their own language.
It is up to all of us who teach writing of any kind to come back to this statement with our students. Show them that their voices matter by allowing them to write in their own beautiful words. These words might not always be perfect. There will most likely be room for improvement with grammar, punctuation, and spelling. After all, writing is not an or communication style, and it takes time to learn how to do it well. For some students, taking the time to learn is perhaps less important to them than learning other concepts that they feel are now more important.
Looking ahead
I have a few thoughts about where writing generally, and specifically in K-20, might (or maybe should) be headed.
1) AI detection tools will hopefully go away, as will this need for humans to “catch” others in their gen AI use for writing. Speaking as someone with ADHD and possible/probable ASD, there are times when gen AI really helps me find and define my thoughts more clearly. The ideas are mine, but the words are like alphabet letters in a bowl of soup — floating and hard to catch. Lance Cummings also posts about the ways that gen AI tools help him with his ADHD writing needs such as with . There are also times when my big vocabulary and sometimes odd way of approaching a topic make me sound like I used gen AI when I didn’t.
2) We can and should still maintain high standards for writing and academic integrity. But before anyone accusing me of allowing a cheating free-for-all in my classrooms, let me point you to work by Sarah Elaine Eaton on or the recent book by Tricia Bertam Gallant and David A. Rettinger called “”. Those ideas, though, are not mutually exclusive from letting students use the words that come naturally to them when they write.
3) Writer voices matter. This is as much a business statement as a writing statement. I’m putting on my tech writing instructor hat for a minute to point out that our audiences care about what we say and how we say things. If I see an ad for a product and it stands out because of how the copy is written, that’s fun. But if every ad copy starts to sound the same, then I tune those words out. As a consumer, I am less likely to remember a jingle or a commentary if it sounds like all the others.
To make my point, if you are old enough to remember when jingles used to be in ads, think of one that has stuck with you — “Meow, meow, meow, meow. Meow, meow, meow, meow….” Cat Chow anyone? Or “The thing about a Bonbon, it’s almost always gone gone.” And now I want Bonbons ice cream treats. Those memorable words, and often music, stick with and we remember those products when we go to the store.
4) Writing still matters. People are still reading, and they usually want to read well-written and thoughtful pieces. Even if you prefer genres that may be considered less serious such as romance or young adult fiction, the odds are that you want to enjoy the reading process. Writing isn’t just about transmission of information. It can also take us on journeys, teach us new things, and help us better understand complex ideas. Readers want good writing.
For now
In this moment, maybe we can step back and think about what matters to us as humans. While it is really cool that <insert program name> can now beat whatever new human metric we have set as the bar, is that what matters to us? Possibly, and also maybe not. What I keep seeing people come back to is the idea of what makes us human and how gen AI can’t yet (and maybe will never) completely replace those parts of humanity. Our language, while replicable by gen AI, still has specific meaning to the user and the reader/listener.
The words that we choose as our communication pathway may matter (imagine MLK’s “I have a dream” speech being written with different words that lack the musicality and cadence of the original), or they may not matter as much such as when telling a story about your day. (And this gets into an understanding of rhetoric and why we use the words that we use. I’ll save that for another conversation.) But our ideas still have value.
At the end of the day, my teaching mission remains twofold. 1) to help my students learn to communicate their ideas, and 2) to help my students reach their audiences where those audiences are at. My children understand when I say “no cap” if I’m trying to communicate a truth to them, but I probably wouldn’t use that in an earnings report. So let our students write. Let them play with language. Let them learn its power and help them find their own. They might choose to use gen AI tools to generate text here and there, which is where the discussions of ethics should come in, but maybe we can also help them learn to love their own voices and word choices, just as they are.
—
*The only AI generated thing here is the graphic from ChatGPT. The rest is me, in all my humanness and fallibility.
*Word really hates my hedges (a rhetorical tool for understating an issue). But letting it edit them out changes my meaning and my tone. Instead, I choose to let my voice and my hesitancy shine through.
