AI is all so pervasive now, but I would like to ask those that use it, "Don't you feel like you are cheating the authors that unknowingly contributed their best words to it?"
They don't care. It's sad, but it's true. They don't think there's anything wrong with it at all. Worse, they're putting their name on it as if they did something more than a Google search 😞
Not just cheating the writers whose work has been used without their permission, but writers who spend hours and hours carefully crafting their work, actually doing the work.
I use Grammarly’s free version and only allow it to add commas, which I sometimes refuse to use. I want my writing to sound like me, not some damn robot. It quite often suggests I remove words for brevity. No, and crap no. I write like I write and won’t change it to sound like someone else.
Same. I use ProWritingAid for the punctuation help but don't pay any attention to the "take out __ for brevity or readability." No, thanks. I like my words the way I write them, thank you very much.
I use ProWriting as well for spelling and punctuation and go over my work. I do not want generative AI (ChatGBT), but Assistive AI is Grammarly and Prowriting...with the caveat that you don't use the beta reader or manuscript editor or whatever to rewrite your document...NO thank you. Also, because I don't trust it to not use my words to train AI.
Same here. I think about my words usage, thanks. I don’t need that kind of editing… usually it takes something I’ve written and makes it bland and changes the whole feel of the thing… frustrating as hell…
how do I turn off copilot which is wasting so much of my time incorrectly editing my sentences as I write them?
This is the main reason that I stopped using ChatGPT, along with the environmental impact. I also think that I'll get better at writing by putting in the effort to plan, practice, and edit on my own without any input from a robot.
That is a big one, yes. And not just energy. So much water fresh water and it's the world's scarcest resource. 2% of the world's water is usable and we're evaporating it in Olympic swimming pool quantities every day.
A few people have told me they ask for feedback, not correction. Like, instead of letting ChatGPT just write for them, they ask where it's weak. But that doesn't seem to be the most common use.
Question. Is there anything to stop AI from literally plagiarizing full paragraphs or sentences? I am guessing it already does and there's nothing in place to stop it. Besides the missing humanity, the heart and soul of writing in my world, I really don't like the fact that no one can tell if it is AI or not, same with images. We are being drawn into a totally fake world and smiling and laughing as we go.
The programmers. When it first launched, it would return full articles that were behind a paywall. That was the basis of one of the first lawsuits. So they did an "update" to tell it not to return full pieces. I don't know what the current maximum is for text repeated verbatim
Fascinating. Really helpful because I now understand a lot more about how AI actually works - and yes, now I'm even more determined never ever to use it. Thank you!
I sometimes feel guilty when I use it to create examples, like I did in this post. But I'm not using it every day, or passing it off as my own writing. So I guess there's that lol
These stats are disturbing. I'm not comfortable engaging with ChatGPT for anything, and cannot fathom asking it to spit out entire essays or articles. And to think that people are earning money from this (stolen!) work.
I, too, struggle to accept my writing. I also feel the same about my photography. For every image I am comfortable sharing, there are dozens to hundreds of discarded ones. AI is also stealing from the work of photographers. I am beginning to value the imperfections in my writing and photographs. If flaws are proud markers of honest work, then here's to embracing imperfection, both real and perceived.
God, I hear you on the discarded work. At any given time, I have close to 100 drafts that I never published because I think they're garbage, not good enough. Sometimes I go through them, see if I can fix up something instead of starting from scratch. And I have folders and folders of photography no one has ever seen except my kid and my brother and they both say it's really good, but ugh, that feeling.
As a teacher I get ai to spit out everything from lists to worksheets to essays. It saves me a lot of time, but I use it for mostly low level learners. The style is pretty obvious - I call it relentless obviousness. "Echoes of laughter," "ghosts of games," from your examples are not obvious in of themselves, but the repetition of those kinds of descriptions, and also collocations, is definitely a tell-tale sign. Also, I'd look out for what I call mechanically-correct dependent clauses! After a while, you'll see ai prose is rife with them. So, in my opinion, those who are using ai to "write" stories are just fooling themselves. Thanks for the article!
I can definitely see a use case for things like making multiple choice quizzes or worksheets based on info you (general sense) have fed it. Sort of like Kahoot or steroids.
Cheers Kevin, for my own writing I use it for discussions. For example just had a long back and forth with Deep Seek about how abstract nouns are the foundation of theme. Then my typical prompt will be: "well how many types or categories of abstract nouns are there?" And it answered - - concept, quality, and state of being - - so yes its a great tool!
Hi Linda, thank you too! I think ai is a wonderful tool to investigate your own ideas on a deeper level, but ironically, the very thing that makes it good at writing formal documents: like a parking ticket, lol, is what hinders it from writing good prose. Cheers
That’s the same reason I won’t use AI for writing.
The 2nd reason not to use AI is when you use AI for writing, you do not develop the writing muscles and depth of craft.
You simply learn to parrot AI.
I remember that being part of the conversation in your previous post as well. And I genuinely think at this point, which may change with new information, that those who write with AI are more interested in the output, the clicks, the likes, the money, the comments, than they are enjoying the process.
I loved your painting example.
If you don’t enjoy painting so much, then why not just buy art instead.
Different priorities are valid. But I can’t reconcile making money off stolen work.
I believe it was you who said, fine to call themselves writers, just don’t try to tell us it’s the same thing. It’s not.
Craft writers— it’s not gatekeeping and it’s not elitist. It’s simply the people who actually enjoy the process, who often see the journey is more important than the destination.
Thank you, Leo. How bizarre that we have to even have these conversations, you know? The work was stolen, used without permission. Why is that even allowed?
There’s something badass about not using AI, like you’re still willing to get into the trenches. I also don’t use it for my Substack writing. I like to think us old fashioned people need to stay strong for the younger generation who doesn’t remember a time before AI.
When my young kids are old enough to start creative writing, I will be a (probably irritating) inspiration for them to do it the right way :)
I love that Gabrielle. It is kind of badass to get in the trenches and I love that you're going to encourage your kids to do it that way. Best thing for them, ever. Especially when they're young and their imaginations are so active!!
Being a senior citizen, I know what AI is but do not fully understand the complete concept. Being a voracious reader (I joined my first book club 70 years ago at age 5), I am disappointed and astonished at the percentage of writers who use AI. I doubt I could identify it (if the thread throughout the piece didn’t make sense or the plot didn’t jive, I’d assume it was an inferior piece of writing). With all of the lies and grifting and deceit in the world today, this is just one more thing that makes me sad. Thank you, Linda, for staying true to your craft!
Thank you, Barbara. Know what the biggest sign of AI is? I heard one educator call it surface profundity. Meaning it feels profound at first blush, but there's no depth or personal experience in it. Once I read that, it was an aha moment for me. And thank you, too!
Good article, Linda. I’m getting better at spotting AI generated text the more I read it. OMG, I’m becoming an algorithm! 😳😊. Just kidding. But I find most AI text gets “every word, phrase, description sounding almost perfect. Or perhaps a better way of putting it - without obvious flaws or shortcomings, which is very human. It’s like the writer had the means and time to find just the right words. Which it did, after consuming ten billion words of the best writers on the planet. Normal writers won’t spend two hours on one sentence, it would drive us crazy. But AI can and does , all within a second or two. But because 82% of all humans are having their attention span and critical thinking reduced daily we can’t see the difference and more importantly don’t care. These are the same humans consuming ultra processed foods with enough sugar to sweeten Lake Erie. And we know this - but it’s fast and easy and we don’t have to waste time cooking with there’s so much Tik Tok videos to look at. I think most humans with 2025 thinking capacity can’t spot AI. But many humans with 20th Century thinking can. My opinion. Cheers. 😊
This rings true for me. I write in blocks and chunks. (highly technical term). Narrative nonfiction. I have a particular voice and vivid detail only I know. Just for fun, I had chat gpt write and interlude. Fed it basic facts of the events. It was horrible. I ended up trashing that and. Writing the whole thing on my own. Sometimes I will load my blocks and chunks into gpt project memory. I have organized into what I think is a logical sequence and ask gpt to evaluate if this is the best organization of the text. And the feedback I get is a choice. Option A, Option B. I decide what makes sense to me and I make the change. Sometimes I ask if there is anything out there with my same words word order etc. basically looking for anything I might have inadvertently copied (hasn’t happened yet). I look to my own life for my ideas, so I haven’t felt I’ve been using AI to write. I get the words on the page. I do ask for editing suggestions. I don’t know. Now I am rethinking the whole process. 🤷♀️
You made me laugh, Jean. I have not read the blocks and chunks, of course, but I'm willing to bet your own words are way more powerful than anything AI can add. Because it can write smoothly, but it has so little depth!
Kinda seriously though, how is asking for editing feedback from AI different from working with humans editor? Ok logistics aside, theoretically, where is the line there? Because if I have to rely on my Word skills, I will be long dead before I’ll be able to publish anything. And Word of giving some editing feedback anyway. And it has a thesaurus. Just sayin’
“how is asking for editing feedback from AI different from working with humans editor?”
Such a good question, one I’ve been pondering a lot myself bcs of the work I do in this field and in the adjacent fields of content coaching and language-teaching.
On the one hand, there are no easy answers to this question, right? Bcs it really can be helpful in so many arenas to have tools like spell-check and grammar-check. And those are especially helpful for my students with disabilities in this area (one with major dyslexia worked extra hard on his spelling struggles, at the same time we all realized that he would always need extra help in this area). And I know it’s a similar case with English-language adopters trying to make sure they can get their message across clearly with less of a language barrier.
Yet on the other hand, there are so many times when even spell check and grammar check get it wrong. Just within the past week, I’ve had two instances where the current person I work with asked “why does the program keep underlining this word?” And when I looked at the word, I could see a) it was actually used correctly and well, b) why/how the word had confused the program into thinking it was wrong. It was a good reminder to him that the program isn’t the decider of what is right/correct.
And I think that’s at least part of the difference between machine-based editing and human editing. Though I am not sure I can name that exact quality yet. Maybe “discernment”? (Which you pointed out in one of your earlier comments, too.)
The machine misses nuance that a human can pick up. The machine is comparing your patterns with its predicted models and misses the less-predictable but fully-correct (and unique!) ways you express things. And sometimes originality obscures clarity, right? But sometimes it brings the clarity, and the machine can’t tell you how it will land with your reader like a human editor can (and even human editors can’t predict how every reader will take it: look at books that have garnered posthumous fame in a much later generation, right?)
The machine also can’t ask probing questions about your meaning to help you find the better words for what you are trying to say — quite a large number of my editing sessions with someone will hinge on questions like these that reveal a whole different collection of words or phrases than the writer began with . . . simply because the questions I asked brought greater clarity (for both of us).
And even with this kind of editing process, I have often asked myself, “if I suggest a different phrasing, has this now become my work as well? Is it less authentically hers/his?” In fact, this is a really important question when I’m working with high school/college students. And this question (of regulatory ethics) is also a key difference between machine editing and human editing: the machine can’t ask this of itself (and bcs it’s a machine, a tool, we often don’t).
The answer to my self-regulatory question is never black-and-white, usually a scale — depending on factors like what the paper is for (if it’s a school piece, it’s important that I balance helping the student find clarity while making sure she expresses it with words that show her actual level of proficiency; if it’s for a friend’s resume, the goals are often quite different).
And this human factor, too—the borrowing of words directly from another human—is why writers list their editors in an Acknowledgments page, right? Bcs the editors midwife the book into being the best thing the author envisioned.
As long-winded as this comment is, these are still just seed thoughts, of course, not definitive. I think that one of the things we are grappling with when it comes to AI tools (as they exist today) is scale. At what point does a tool cease being an affordable/time-saving substitute for the human editor? at what point does it do more harm than it does good?
Thank you for your thoughtful answer—there are so many considerations when using AI. One thing I appreciate about AI is the ability to ask qualitative questions, like determining whether a passage is redundant or whether it should be a full chapter or an interlude. With my use of ChatGPT, I have a standing instruction not to make changes or edits unless I explicitly approve them.
For redundancy, AI can highlight similar passages, helping me see which areas could be reworked or removed. For the decision on chapter versus interlude, AI typically gives suggestions. For example, if included as part of a chapter, it might emphasize certain themes, while using it as an interlude could provide a break for the reader, offering a lighter or more humorous tone. These are questions I'd typically discuss with an editor, but AI provides suggestions that I can choose to pursue or ignore.
Another helpful feature is revising a few lines or a paragraph and asking AI to show how it fits with the rest of the piece. It effectively reflects back what I've already typed, allowing me to see how changes integrate with the entire work.
This capability makes AI a useful tool in the creative process, offering insights and perspectives while keeping the final decision-making in my hands.
Thank you for explaining even more of your process to me. I see how you’re using it with a similar kind of intentionality that you would use with other writing tools and with a human editor.
The key difference I see is that, as the sole human in the interaction, you’re the one who must come up with all the questions to ask.
It simulates collaboration and yet is also not collaboration . . . fascinating!
Thanks, Joe. I agree with the 20th century thinking, for sure. Know what kills me? The don't care part. I can't even imagine not caring but maybe that's because the craft of writing means so much to me :)
I feel the same way. Would only use AI to help me write but wouldn’t use it to actually write my stories. I’d rather deal with imposter phenomena than to have AI write for me.
Hi, Linda! I'm a new subscriber, and your newsletter is one I look forward to every time. Thanks for doing battle with your self-doubt, because I value your work and your transparency, your voice and your writing! I'm looking forward to reading more from you. Keep it up!
Thank you for the article as always you break things down really well. I would say that the argument that the developers have stolen copyrighted work is more nuanced and authors and artists should consider using a stronger argument. I discuss this in an article I wrote on some recent court decisions on copyright infringement (not posting link cause I think that’s tacky but you can find it if you click on my name). The court says that blatantly pirated work that was fed into the LLM was infringement but the court did not agree that works the AI companies purchased and fed to the LLM was also infringement. So my point is the argument can’t just be about piracy it needs to focus on AI being so outside anything we could have predicted that we need new specific laws that deal with how AI can be trained. The argument needs to be that as AI gets better and better it will stifle original authors ability to make a living. The argument also should be that AI companies should have to get a special license to “adapt” a copyrighted work just like movie studios do because if we just rely on the copyright law as written it seems the court is saying that if they purchased the book then all is permitted.
The genie is out of the bottle. We were told that social media was a tool for connection, that it would benefit us. Instead, it has become a tool of mostly destruction. And the tech bros are making a fortune off of our addiction to social media. Teens are dying by suicide because of the power of social media. We shrug our shoulders. Now AI has been unleashed, and we shrug our shoulders. Nothing is free in a capitalist economy, so the hidden costs are a dulling of the masses and a data trail for authoritarians to access. Did you ask AI about sanctuary cities out of curiosity? Did you ask ChatGPT about self-induced abortions? Did you ask AI how effective protests are? All searchable ideas, and searchable by those who want to squash freedoms.
Well, that's jolly, isn't it? Enough to make me want to chuck the whole thing in. My only saving grace is that I never think my writing is crap. I always like it. If I don't like it, I fix it until I do. Maybe that makes me an oddball, but if I agonized over everything I write, I wouldn't bother.
You made me laugh, Sue. I do agonize, but I also can't not write. So I just deal with it and move on. I'm glad you don't struggle with that, nice that we don't all have to lol
AI is all so pervasive now, but I would like to ask those that use it, "Don't you feel like you are cheating the authors that unknowingly contributed their best words to it?"
From the comments I get, they don't seem to lol
Ugh... I am so out of step...
I don't think you're the one out of step :)
They don't care. It's sad, but it's true. They don't think there's anything wrong with it at all. Worse, they're putting their name on it as if they did something more than a Google search 😞
Yeah. That's the part gets me, too
Not just cheating the writers whose work has been used without their permission, but writers who spend hours and hours carefully crafting their work, actually doing the work.
I use Grammarly’s free version and only allow it to add commas, which I sometimes refuse to use. I want my writing to sound like me, not some damn robot. It quite often suggests I remove words for brevity. No, and crap no. I write like I write and won’t change it to sound like someone else.
Same. I use ProWritingAid for the punctuation help but don't pay any attention to the "take out __ for brevity or readability." No, thanks. I like my words the way I write them, thank you very much.
Bravo, kindred spirit!
I use ProWriting as well for spelling and punctuation and go over my work. I do not want generative AI (ChatGBT), but Assistive AI is Grammarly and Prowriting...with the caveat that you don't use the beta reader or manuscript editor or whatever to rewrite your document...NO thank you. Also, because I don't trust it to not use my words to train AI.
Same here. I think about my words usage, thanks. I don’t need that kind of editing… usually it takes something I’ve written and makes it bland and changes the whole feel of the thing… frustrating as hell…
how do I turn off copilot which is wasting so much of my time incorrectly editing my sentences as I write them?
Is copilot part of Grammarly or Word? I usually Google this kind of question or go to the app’s website for help.
Sorry, it’s not part of Grammarly, but Word.
This is the main reason that I stopped using ChatGPT, along with the environmental impact. I also think that I'll get better at writing by putting in the effort to plan, practice, and edit on my own without any input from a robot.
The environmental impact, yes! Yet another (and the biggest, in the grand scheme of things) reason to not use AI.
That is a big one, yes. And not just energy. So much water fresh water and it's the world's scarcest resource. 2% of the world's water is usable and we're evaporating it in Olympic swimming pool quantities every day.
A few people have told me they ask for feedback, not correction. Like, instead of letting ChatGPT just write for them, they ask where it's weak. But that doesn't seem to be the most common use.
ChatGpt bad quality poems
Question. Is there anything to stop AI from literally plagiarizing full paragraphs or sentences? I am guessing it already does and there's nothing in place to stop it. Besides the missing humanity, the heart and soul of writing in my world, I really don't like the fact that no one can tell if it is AI or not, same with images. We are being drawn into a totally fake world and smiling and laughing as we go.
The programmers. When it first launched, it would return full articles that were behind a paywall. That was the basis of one of the first lawsuits. So they did an "update" to tell it not to return full pieces. I don't know what the current maximum is for text repeated verbatim
Thank you.
Fascinating. Really helpful because I now understand a lot more about how AI actually works - and yes, now I'm even more determined never ever to use it. Thank you!
I sometimes feel guilty when I use it to create examples, like I did in this post. But I'm not using it every day, or passing it off as my own writing. So I guess there's that lol
These stats are disturbing. I'm not comfortable engaging with ChatGPT for anything, and cannot fathom asking it to spit out entire essays or articles. And to think that people are earning money from this (stolen!) work.
I, too, struggle to accept my writing. I also feel the same about my photography. For every image I am comfortable sharing, there are dozens to hundreds of discarded ones. AI is also stealing from the work of photographers. I am beginning to value the imperfections in my writing and photographs. If flaws are proud markers of honest work, then here's to embracing imperfection, both real and perceived.
God, I hear you on the discarded work. At any given time, I have close to 100 drafts that I never published because I think they're garbage, not good enough. Sometimes I go through them, see if I can fix up something instead of starting from scratch. And I have folders and folders of photography no one has ever seen except my kid and my brother and they both say it's really good, but ugh, that feeling.
As a teacher I get ai to spit out everything from lists to worksheets to essays. It saves me a lot of time, but I use it for mostly low level learners. The style is pretty obvious - I call it relentless obviousness. "Echoes of laughter," "ghosts of games," from your examples are not obvious in of themselves, but the repetition of those kinds of descriptions, and also collocations, is definitely a tell-tale sign. Also, I'd look out for what I call mechanically-correct dependent clauses! After a while, you'll see ai prose is rife with them. So, in my opinion, those who are using ai to "write" stories are just fooling themselves. Thanks for the article!
I can definitely see a use case for things like making multiple choice quizzes or worksheets based on info you (general sense) have fed it. Sort of like Kahoot or steroids.
Cheers Kevin, for my own writing I use it for discussions. For example just had a long back and forth with Deep Seek about how abstract nouns are the foundation of theme. Then my typical prompt will be: "well how many types or categories of abstract nouns are there?" And it answered - - concept, quality, and state of being - - so yes its a great tool!
Right? I love that. Because it can dig for information really well.
Oh I love that. You gave me language for some of what stands out, but I hadn't really thought about why. Thank you!
Hi Linda, thank you too! I think ai is a wonderful tool to investigate your own ideas on a deeper level, but ironically, the very thing that makes it good at writing formal documents: like a parking ticket, lol, is what hinders it from writing good prose. Cheers
I so agree. It chooses words based on probability, whereas good writers find ways to use words in new and novel ways.
That’s the same reason I won’t use AI for writing.
The 2nd reason not to use AI is when you use AI for writing, you do not develop the writing muscles and depth of craft.
You simply learn to parrot AI.
I remember that being part of the conversation in your previous post as well. And I genuinely think at this point, which may change with new information, that those who write with AI are more interested in the output, the clicks, the likes, the money, the comments, than they are enjoying the process.
I loved your painting example.
If you don’t enjoy painting so much, then why not just buy art instead.
Different priorities are valid. But I can’t reconcile making money off stolen work.
I believe it was you who said, fine to call themselves writers, just don’t try to tell us it’s the same thing. It’s not.
Craft writers— it’s not gatekeeping and it’s not elitist. It’s simply the people who actually enjoy the process, who often see the journey is more important than the destination.
Loved your article. 💕
Thank you, Leo. How bizarre that we have to even have these conversations, you know? The work was stolen, used without permission. Why is that even allowed?
There’s something badass about not using AI, like you’re still willing to get into the trenches. I also don’t use it for my Substack writing. I like to think us old fashioned people need to stay strong for the younger generation who doesn’t remember a time before AI.
When my young kids are old enough to start creative writing, I will be a (probably irritating) inspiration for them to do it the right way :)
I love that Gabrielle. It is kind of badass to get in the trenches and I love that you're going to encourage your kids to do it that way. Best thing for them, ever. Especially when they're young and their imaginations are so active!!
Being a senior citizen, I know what AI is but do not fully understand the complete concept. Being a voracious reader (I joined my first book club 70 years ago at age 5), I am disappointed and astonished at the percentage of writers who use AI. I doubt I could identify it (if the thread throughout the piece didn’t make sense or the plot didn’t jive, I’d assume it was an inferior piece of writing). With all of the lies and grifting and deceit in the world today, this is just one more thing that makes me sad. Thank you, Linda, for staying true to your craft!
Thank you, Barbara. Know what the biggest sign of AI is? I heard one educator call it surface profundity. Meaning it feels profound at first blush, but there's no depth or personal experience in it. Once I read that, it was an aha moment for me. And thank you, too!
Good article, Linda. I’m getting better at spotting AI generated text the more I read it. OMG, I’m becoming an algorithm! 😳😊. Just kidding. But I find most AI text gets “every word, phrase, description sounding almost perfect. Or perhaps a better way of putting it - without obvious flaws or shortcomings, which is very human. It’s like the writer had the means and time to find just the right words. Which it did, after consuming ten billion words of the best writers on the planet. Normal writers won’t spend two hours on one sentence, it would drive us crazy. But AI can and does , all within a second or two. But because 82% of all humans are having their attention span and critical thinking reduced daily we can’t see the difference and more importantly don’t care. These are the same humans consuming ultra processed foods with enough sugar to sweeten Lake Erie. And we know this - but it’s fast and easy and we don’t have to waste time cooking with there’s so much Tik Tok videos to look at. I think most humans with 2025 thinking capacity can’t spot AI. But many humans with 20th Century thinking can. My opinion. Cheers. 😊
This rings true for me. I write in blocks and chunks. (highly technical term). Narrative nonfiction. I have a particular voice and vivid detail only I know. Just for fun, I had chat gpt write and interlude. Fed it basic facts of the events. It was horrible. I ended up trashing that and. Writing the whole thing on my own. Sometimes I will load my blocks and chunks into gpt project memory. I have organized into what I think is a logical sequence and ask gpt to evaluate if this is the best organization of the text. And the feedback I get is a choice. Option A, Option B. I decide what makes sense to me and I make the change. Sometimes I ask if there is anything out there with my same words word order etc. basically looking for anything I might have inadvertently copied (hasn’t happened yet). I look to my own life for my ideas, so I haven’t felt I’ve been using AI to write. I get the words on the page. I do ask for editing suggestions. I don’t know. Now I am rethinking the whole process. 🤷♀️
You made me laugh, Jean. I have not read the blocks and chunks, of course, but I'm willing to bet your own words are way more powerful than anything AI can add. Because it can write smoothly, but it has so little depth!
Like some bad dates I’ve had. 😝
Kinda seriously though, how is asking for editing feedback from AI different from working with humans editor? Ok logistics aside, theoretically, where is the line there? Because if I have to rely on my Word skills, I will be long dead before I’ll be able to publish anything. And Word of giving some editing feedback anyway. And it has a thesaurus. Just sayin’
Jean
“how is asking for editing feedback from AI different from working with humans editor?”
Such a good question, one I’ve been pondering a lot myself bcs of the work I do in this field and in the adjacent fields of content coaching and language-teaching.
On the one hand, there are no easy answers to this question, right? Bcs it really can be helpful in so many arenas to have tools like spell-check and grammar-check. And those are especially helpful for my students with disabilities in this area (one with major dyslexia worked extra hard on his spelling struggles, at the same time we all realized that he would always need extra help in this area). And I know it’s a similar case with English-language adopters trying to make sure they can get their message across clearly with less of a language barrier.
Yet on the other hand, there are so many times when even spell check and grammar check get it wrong. Just within the past week, I’ve had two instances where the current person I work with asked “why does the program keep underlining this word?” And when I looked at the word, I could see a) it was actually used correctly and well, b) why/how the word had confused the program into thinking it was wrong. It was a good reminder to him that the program isn’t the decider of what is right/correct.
And I think that’s at least part of the difference between machine-based editing and human editing. Though I am not sure I can name that exact quality yet. Maybe “discernment”? (Which you pointed out in one of your earlier comments, too.)
The machine misses nuance that a human can pick up. The machine is comparing your patterns with its predicted models and misses the less-predictable but fully-correct (and unique!) ways you express things. And sometimes originality obscures clarity, right? But sometimes it brings the clarity, and the machine can’t tell you how it will land with your reader like a human editor can (and even human editors can’t predict how every reader will take it: look at books that have garnered posthumous fame in a much later generation, right?)
The machine also can’t ask probing questions about your meaning to help you find the better words for what you are trying to say — quite a large number of my editing sessions with someone will hinge on questions like these that reveal a whole different collection of words or phrases than the writer began with . . . simply because the questions I asked brought greater clarity (for both of us).
And even with this kind of editing process, I have often asked myself, “if I suggest a different phrasing, has this now become my work as well? Is it less authentically hers/his?” In fact, this is a really important question when I’m working with high school/college students. And this question (of regulatory ethics) is also a key difference between machine editing and human editing: the machine can’t ask this of itself (and bcs it’s a machine, a tool, we often don’t).
The answer to my self-regulatory question is never black-and-white, usually a scale — depending on factors like what the paper is for (if it’s a school piece, it’s important that I balance helping the student find clarity while making sure she expresses it with words that show her actual level of proficiency; if it’s for a friend’s resume, the goals are often quite different).
And this human factor, too—the borrowing of words directly from another human—is why writers list their editors in an Acknowledgments page, right? Bcs the editors midwife the book into being the best thing the author envisioned.
As long-winded as this comment is, these are still just seed thoughts, of course, not definitive. I think that one of the things we are grappling with when it comes to AI tools (as they exist today) is scale. At what point does a tool cease being an affordable/time-saving substitute for the human editor? at what point does it do more harm than it does good?
Thank you for your thoughtful answer—there are so many considerations when using AI. One thing I appreciate about AI is the ability to ask qualitative questions, like determining whether a passage is redundant or whether it should be a full chapter or an interlude. With my use of ChatGPT, I have a standing instruction not to make changes or edits unless I explicitly approve them.
For redundancy, AI can highlight similar passages, helping me see which areas could be reworked or removed. For the decision on chapter versus interlude, AI typically gives suggestions. For example, if included as part of a chapter, it might emphasize certain themes, while using it as an interlude could provide a break for the reader, offering a lighter or more humorous tone. These are questions I'd typically discuss with an editor, but AI provides suggestions that I can choose to pursue or ignore.
Another helpful feature is revising a few lines or a paragraph and asking AI to show how it fits with the rest of the piece. It effectively reflects back what I've already typed, allowing me to see how changes integrate with the entire work.
This capability makes AI a useful tool in the creative process, offering insights and perspectives while keeping the final decision-making in my hands.
Thank you for explaining even more of your process to me. I see how you’re using it with a similar kind of intentionality that you would use with other writing tools and with a human editor.
The key difference I see is that, as the sole human in the interaction, you’re the one who must come up with all the questions to ask.
It simulates collaboration and yet is also not collaboration . . . fascinating!
Thanks, Joe. I agree with the 20th century thinking, for sure. Know what kills me? The don't care part. I can't even imagine not caring but maybe that's because the craft of writing means so much to me :)
I feel the same way. Would only use AI to help me write but wouldn’t use it to actually write my stories. I’d rather deal with imposter phenomena than to have AI write for me.
I so agree, Rosanne. I battle feeling like a sham all the time, but I keep on doing it because it feels honest. :)
Hi, Linda! I'm a new subscriber, and your newsletter is one I look forward to every time. Thanks for doing battle with your self-doubt, because I value your work and your transparency, your voice and your writing! I'm looking forward to reading more from you. Keep it up!
Aww, Kate, thank you. What a truly nice thing to say, and I appreciate it a ton! :)
Thank you for the article as always you break things down really well. I would say that the argument that the developers have stolen copyrighted work is more nuanced and authors and artists should consider using a stronger argument. I discuss this in an article I wrote on some recent court decisions on copyright infringement (not posting link cause I think that’s tacky but you can find it if you click on my name). The court says that blatantly pirated work that was fed into the LLM was infringement but the court did not agree that works the AI companies purchased and fed to the LLM was also infringement. So my point is the argument can’t just be about piracy it needs to focus on AI being so outside anything we could have predicted that we need new specific laws that deal with how AI can be trained. The argument needs to be that as AI gets better and better it will stifle original authors ability to make a living. The argument also should be that AI companies should have to get a special license to “adapt” a copyrighted work just like movie studios do because if we just rely on the copyright law as written it seems the court is saying that if they purchased the book then all is permitted.
Interesting. Things that make you go hmmmm. I use AI for research. Period.
The genie is out of the bottle. We were told that social media was a tool for connection, that it would benefit us. Instead, it has become a tool of mostly destruction. And the tech bros are making a fortune off of our addiction to social media. Teens are dying by suicide because of the power of social media. We shrug our shoulders. Now AI has been unleashed, and we shrug our shoulders. Nothing is free in a capitalist economy, so the hidden costs are a dulling of the masses and a data trail for authoritarians to access. Did you ask AI about sanctuary cities out of curiosity? Did you ask ChatGPT about self-induced abortions? Did you ask AI how effective protests are? All searchable ideas, and searchable by those who want to squash freedoms.
Oh God, Sue. So much. Pandora's box has been opened.
Well, that's jolly, isn't it? Enough to make me want to chuck the whole thing in. My only saving grace is that I never think my writing is crap. I always like it. If I don't like it, I fix it until I do. Maybe that makes me an oddball, but if I agonized over everything I write, I wouldn't bother.
You made me laugh, Sue. I do agonize, but I also can't not write. So I just deal with it and move on. I'm glad you don't struggle with that, nice that we don't all have to lol