This is a very strong and uncomfortable piece, in the best sense.
What lands for me is the connection between writing and photography. In both, the real value is not polish. It is contact. A photograph matters because someone stood somewhere, noticed something, waited, chose, failed, tried again. Writing has a similar burden. It should carry the pressure of an actual mind meeting actual experience.
The problem with too much AI-assisted prose is not simply that it sounds “artificial.” It is that it often removes the friction where meaning lives. The odd detail, the clumsy first attempt, the sentence that has to be fought into shape, the specific street, room, weather, person, or year. Those are not decorative details. They are evidence.
Your point about category lists standing in for memory feels especially important. “Airports, diners, deserts, motels” can create the mood of experience while avoiding the risk of naming the actual place. But the truth is usually in the thing that does not generalize neatly: the bad coffee, the green plastic chair, the woman who would not look at the camera, the rain that changed the whole frame.
I also appreciate that you are not only arguing against AI. You are arguing for practice. Reading aloud. Revising. Developing an ear. Letting syntax become part of authorship rather than outsourcing the discomfort of learning it.
For photographers and creatives, this matters because the written frame around the work is increasingly part of the work. Captions, essays, artist statements, newsletters, books: these things shape trust. If the language becomes generic, the pictures are asked to carry not only meaning, but credibility too.
The real recommendation here may be simple and demanding: write closer to the contact sheet.
I see the same basic outline on most writing these days. I even see it in the NYT. Liberals and conservatives alike use AI trying to fill in every open space with meaningless crap. And I’m shocked that people read it, and then comment on it. Several people have told me how they disagreed, then read my examples and realized they had read this guy’s stuff. Why isn’t this the same as using AI to create a photo through a few prompts? Got me. But soon no one will be able to think.
Well thought out! As a photographer, reader, writer, and copyeditor I'm glad to see your detailed breakdown of some of AI's failings. I suspect the fight against AI slop will be long, though. But count me in for the duration.
What you say makes absolute sense … when you write in your native language. I will not hide that I use AI assistance because my English is not perfect. You might say: why not write in French and use AI only to translate? Unfortunately, my English is good enough to spot bad translations and question the resulting draft. Correcting an AI-translated text when you aim for a certain quality of style and accuracy is extremely time-consuming, and I am often not fully satisfied with the result.
So I tested the opposite approach: write directly in English, and have AI correct the occasional grammatical or lexical mistake, and sometimes suggest a better sentence structure or rhythm. Then, for the sake of efficiency, if I want the same text in French, I ask AI for a first draft translation. Correcting and adjusting in my native language is faster and easier. I don’t need word-for-word accuracy either, and I regularly add, change or cut. Cultural references and insider jokes rarely survive translation anyway.
I use my Apple earpods to do my translating in Japan, German, and Italy, sure. But you're looking for an exception rather than paying attention to the argument. Many people are filling the space up with poor writing, not just poor, but not even their own. I supervise psychologists who use AI to write their clinical evaluations and reports. When we have a clinical meeting, they don't say a word about their reports, or can't defend the diagnosis, because AI wrote it out for them. They know what it is, but can't describe how it was arrived at. This leads to dangerous situations.
I often write back and forth with photographers in Japan; we send our emails translated as well as in our own language, and have time, take time, to listen, get a word right, understand a word that not only is translated, but is difficult because you need cultural knowledge. That's a different situation than filling the space with empty nonsense, long lists of words, one-sentence paragraphs, arguments that don't go anywhere no matter what language you read them in.
If someone wants to put themselves forward as having valuable, original ideas, they need to accept that there are some issues about quality that are important.
I totally share your view and I entirely understand your example. Delegating your thinking to AI to the point where you can no longer defend or explain it is not just lazy, it is a lack of respect for the reader. There is also a meaningful difference between a professional technical use, where AI is a productivity tool, and creative writing, which is your own voice, thoughts and feelings. What I described is closer to having a good editor than to having a ghostwriter.
Jim, thanks for reading that stuff so I don't have to. I think I would just turn off after a couple of paragraphs, even if the subject was supposedly photography, which I am always curious about.
I have enough trouble with the coloured lines that appear under words I put down because the system thinks they could in some ways be clearer or more grammatically precise. The trend encouraged by AI and auto corrections is to a dull, dead language and I am quietly encouraged that there are quite a few people who won't accept it and will continue to write as human beings, warts and all.
Personally, I recommend a diet of Swift, Addison, Keats, Austen and Isherwood (list not supplied by AI) to see what can be done when you harness the English language to your imagination. There are certainly many other examples in many other languages that will confound the machine.
I can agree with a lot of what you wrote, but there are two things I would like to say.
1. I find it funny to use AI to check if someone was using AI for their writing. I just learned a few weeks ago that this is even a thing. But I also learned that AI isn’t always correct. Alex Kustanovic (a writer here on Substack) ran a story he wrote through AI and it said the text was 100% AI. And he did not even use it to check for grammatical errors or the like.
2. I disagree with you about AI can’t teach you how to write. I am not a native English speaker, but I will write in English and then ask AI to check for any grammatical mistakes I have made. Most of these errors I won’t find by reading it again and again, when I don’t know that I made a mistake in the first place. But if “someone” (AI in this case) tells me that I f.e. used the wrong tense of a verb, I get the chance to learn from it.
Yes, as a writing instructor, I use an Ai detector, but if it suspects something, I then pass a paper through two more, and then ask for a narrative description of the issues. That narrative either makes sense or doesn’t. I used to teach a course at NYU on writing for students who, after a first failed semester, obviously needed some help. 20-25 percent did the exercises and did the work; the other 75 percent just didn’t take things seriously. Now, using the narrative description of the problems, we are almost always talking about things that made the work horrid writing. AI or not. If it were any good, I wouldn’t be wasting my time. Still, it is usually pretty obvious when something is either AI or following some pre-digested list of things to do, and they create a monster like Frankenstein’s.
Yes, as a writing instructor, I use an Ai detector, but if it suspects something, I then pass a paper through two more, and then ask for a narrative description of the issues. That narrative either makes sense or doesn’t. I used to teach a course at NYU on writing for students who, after a first failed semester, obviously needed some help. 20-25 percent did the exercises and did the work; the other 75 percent just didn’t take things seriously. Now, using the narrative description of the problems, we are almost always talking about things that made the work horrid writing. AI or not. If it were any good, I wouldn’t be wasting my time. Still, it is usually pretty obvious when something is either AI or following some pre-digested list of things to do, and they create a monster like Frankenstein’s.
I see increasing numbers of Substack "articles" that appear to be bulleted lists without the bullets. Apparently that is also a hallmark of AI? You get a few paragraphs and then the whole article turns into a sequence of one sentence paragraphs. Drives me batty!
I have an MFA in poetry. I was a journalist. I made an adult-living running a content management company. Yet, I read the article you refer to and I did not notice the AI. Only after a fellow Substacker pointed out similar issues that you have pointed out did I notice. I was embarassed. The issues are obvious. But I think a big reason I did not notice was that I was interested in his subject matter. So I wasn't looking for those issues. My bad. Yes, Embarassed! And Maybe I'm getting old and worn out with critiquing every aspect of my life, but I am forgiving of people who don't consider writing to be a strong part of who they are and rely on tech for help. Sure, the writer relied on too much AI. He didn't make it into his own voice. But he was honest about his use of AI. And the tool allowed him to express a concept that was important to him. I can forgive his use of AI for those reasons.
We each have our lines in the sands of technology: presets, anyone? auto-focus and image stabilization, y'all? Auto-spell, children? I say, choose your writing and reading art-creating comfort zone on a case-by-case basis and pick your battles.
Oh, the problem is everywhere, and that guy has a huge following, and everything is AI-generated, the same sentences, essay after essay with a few different words stuck in. My worry is that people who can write will not be able to get jobs, they will go to people who can’t organize a thought, but use Ai to send out their resume.
Like I said, I'm tired of hearing myself critique everything. I am moving to a view of the world in which I only open my mouth to that which i can control, and leave the rest to whatever definition of "progress" makes for IPOs and headlines. i'll look closer at his history of posts and will likely agree with you, since so many others I follow echo your sentiment. In the meantime I'll keep my head down and write the old fashioned way, and let AI generated content cannibalize its own feces. If Substack becomes overwhelmed with robotic thinking and image producing... Well, I'll figure out that step when it's time. The world needs warriors to keep us from straying too far. So thank you. Sincerely. I used to think I was one of them. Not no more.
Editing your own work usually means getting rid of 20-30 percent of what you wrote. Read it out loud while editing, print it, read it, and mark it up. Some AI’s can take your typed original along with your handwritten notes and make corrections for you that you handwrote. I have people I work with do this. You can also ask AI for feedback and then make the corrections yourself. Ask AI if it could ask you 5 questions that, if answered, would make your piece better. Regretfully, many of the “writers” we have here do not make corrections themselves and use AI to do the job for them rather than using it to help them make the changes themselves. Keep writing.
You might be astonished by the myriad Media I’ll peruse & filter through daily as a ‘hunter & gatherer’ - acolyte of Marshall Mcluhan - & ‘generalist
- Flood The Zone With Shit impacts far more than ‘just politics’ - I hired & was hired - over the years within those Media. Translators, Drivers, Makeup Artists, Musicians, Illustrators, Electricians, Programmers, Staging, Production Assistants, Animators, Technicians, Guides, Security, Editors, Wardrobe, Riggers, Typesetters.. & often renting the specialized Gear or / Services required by those experts - for just one day, an hour, even weeks or 15 seconds .. not to mention Talent / Stunt Actors
Those People ? AI is wiping them out daily - We cannot even ‘see them or hear them’ through the Shit Flooding into Social Media 24/7
To err is human. The idea that AI doesn’t err is false. It’s getting better and will continue to do so but I’d feel as though I was cheating to use it in my writing. I know my punctuation isn’t always correct and I did horribly in grammar in school. In the years that followed I found that I enjoyed writing. I’m hoping to improve in my skills but I want it to be all me, even if it takes longer to get my thoughts into words. Appreciate your thoughts here on the subject.
I often include a bibliography of my readings. I supply the details and AI is asked to put it into a specific format. It will remove a book name and replace it, tell me a quate is from another book when I tell it I have the book on my desk in front of me and know who wrote it and when, but didn’t want to write it up in a certain format it says, “Oh, good you caught that. You're right. I asked it to tell me the date that Charle Kirk was shot it told me he wasn’t shot, that I was misremembering, and that I must have some anxiety and should talk to someone. I said I know what I read and was told that people who are upset sometimes make mistakes and misunderstand things. I loaded a copy of the NYT’s for it to read and it said, “Sorry, I hadn’t updated that information.” I asked it what would happen if a 14 year old were told they had basic facts wrong and were emotionally confused if offered no comment.
Ooof. I don’t think I could’ve gotten through that person’s article with these examples you’ve pointed out here. I’ve never used AI myself so it was interesting to see how poorly it edits a text. Yeesh.
This is a very strong and uncomfortable piece, in the best sense.
What lands for me is the connection between writing and photography. In both, the real value is not polish. It is contact. A photograph matters because someone stood somewhere, noticed something, waited, chose, failed, tried again. Writing has a similar burden. It should carry the pressure of an actual mind meeting actual experience.
The problem with too much AI-assisted prose is not simply that it sounds “artificial.” It is that it often removes the friction where meaning lives. The odd detail, the clumsy first attempt, the sentence that has to be fought into shape, the specific street, room, weather, person, or year. Those are not decorative details. They are evidence.
Your point about category lists standing in for memory feels especially important. “Airports, diners, deserts, motels” can create the mood of experience while avoiding the risk of naming the actual place. But the truth is usually in the thing that does not generalize neatly: the bad coffee, the green plastic chair, the woman who would not look at the camera, the rain that changed the whole frame.
I also appreciate that you are not only arguing against AI. You are arguing for practice. Reading aloud. Revising. Developing an ear. Letting syntax become part of authorship rather than outsourcing the discomfort of learning it.
For photographers and creatives, this matters because the written frame around the work is increasingly part of the work. Captions, essays, artist statements, newsletters, books: these things shape trust. If the language becomes generic, the pictures are asked to carry not only meaning, but credibility too.
The real recommendation here may be simple and demanding: write closer to the contact sheet.
I see the same basic outline on most writing these days. I even see it in the NYT. Liberals and conservatives alike use AI trying to fill in every open space with meaningless crap. And I’m shocked that people read it, and then comment on it. Several people have told me how they disagreed, then read my examples and realized they had read this guy’s stuff. Why isn’t this the same as using AI to create a photo through a few prompts? Got me. But soon no one will be able to think.
You are arguing for practice. That’s a notable thing Jim.
Well thought out! As a photographer, reader, writer, and copyeditor I'm glad to see your detailed breakdown of some of AI's failings. I suspect the fight against AI slop will be long, though. But count me in for the duration.
What you say makes absolute sense … when you write in your native language. I will not hide that I use AI assistance because my English is not perfect. You might say: why not write in French and use AI only to translate? Unfortunately, my English is good enough to spot bad translations and question the resulting draft. Correcting an AI-translated text when you aim for a certain quality of style and accuracy is extremely time-consuming, and I am often not fully satisfied with the result.
So I tested the opposite approach: write directly in English, and have AI correct the occasional grammatical or lexical mistake, and sometimes suggest a better sentence structure or rhythm. Then, for the sake of efficiency, if I want the same text in French, I ask AI for a first draft translation. Correcting and adjusting in my native language is faster and easier. I don’t need word-for-word accuracy either, and I regularly add, change or cut. Cultural references and insider jokes rarely survive translation anyway.
I use my Apple earpods to do my translating in Japan, German, and Italy, sure. But you're looking for an exception rather than paying attention to the argument. Many people are filling the space up with poor writing, not just poor, but not even their own. I supervise psychologists who use AI to write their clinical evaluations and reports. When we have a clinical meeting, they don't say a word about their reports, or can't defend the diagnosis, because AI wrote it out for them. They know what it is, but can't describe how it was arrived at. This leads to dangerous situations.
I often write back and forth with photographers in Japan; we send our emails translated as well as in our own language, and have time, take time, to listen, get a word right, understand a word that not only is translated, but is difficult because you need cultural knowledge. That's a different situation than filling the space with empty nonsense, long lists of words, one-sentence paragraphs, arguments that don't go anywhere no matter what language you read them in.
If someone wants to put themselves forward as having valuable, original ideas, they need to accept that there are some issues about quality that are important.
I totally share your view and I entirely understand your example. Delegating your thinking to AI to the point where you can no longer defend or explain it is not just lazy, it is a lack of respect for the reader. There is also a meaningful difference between a professional technical use, where AI is a productivity tool, and creative writing, which is your own voice, thoughts and feelings. What I described is closer to having a good editor than to having a ghostwriter.
Jim, thanks for reading that stuff so I don't have to. I think I would just turn off after a couple of paragraphs, even if the subject was supposedly photography, which I am always curious about.
I have enough trouble with the coloured lines that appear under words I put down because the system thinks they could in some ways be clearer or more grammatically precise. The trend encouraged by AI and auto corrections is to a dull, dead language and I am quietly encouraged that there are quite a few people who won't accept it and will continue to write as human beings, warts and all.
Personally, I recommend a diet of Swift, Addison, Keats, Austen and Isherwood (list not supplied by AI) to see what can be done when you harness the English language to your imagination. There are certainly many other examples in many other languages that will confound the machine.
Let's man the barricades now!
I can agree with a lot of what you wrote, but there are two things I would like to say.
1. I find it funny to use AI to check if someone was using AI for their writing. I just learned a few weeks ago that this is even a thing. But I also learned that AI isn’t always correct. Alex Kustanovic (a writer here on Substack) ran a story he wrote through AI and it said the text was 100% AI. And he did not even use it to check for grammatical errors or the like.
2. I disagree with you about AI can’t teach you how to write. I am not a native English speaker, but I will write in English and then ask AI to check for any grammatical mistakes I have made. Most of these errors I won’t find by reading it again and again, when I don’t know that I made a mistake in the first place. But if “someone” (AI in this case) tells me that I f.e. used the wrong tense of a verb, I get the chance to learn from it.
Here is the substack article I was talking about: https://kustanovich.substack.com/p/when-ai-detector-calls?r=4ylwk&utm_medium=ios
Yes, as a writing instructor, I use an Ai detector, but if it suspects something, I then pass a paper through two more, and then ask for a narrative description of the issues. That narrative either makes sense or doesn’t. I used to teach a course at NYU on writing for students who, after a first failed semester, obviously needed some help. 20-25 percent did the exercises and did the work; the other 75 percent just didn’t take things seriously. Now, using the narrative description of the problems, we are almost always talking about things that made the work horrid writing. AI or not. If it were any good, I wouldn’t be wasting my time. Still, it is usually pretty obvious when something is either AI or following some pre-digested list of things to do, and they create a monster like Frankenstein’s.
Yes, as a writing instructor, I use an Ai detector, but if it suspects something, I then pass a paper through two more, and then ask for a narrative description of the issues. That narrative either makes sense or doesn’t. I used to teach a course at NYU on writing for students who, after a first failed semester, obviously needed some help. 20-25 percent did the exercises and did the work; the other 75 percent just didn’t take things seriously. Now, using the narrative description of the problems, we are almost always talking about things that made the work horrid writing. AI or not. If it were any good, I wouldn’t be wasting my time. Still, it is usually pretty obvious when something is either AI or following some pre-digested list of things to do, and they create a monster like Frankenstein’s.
I see increasing numbers of Substack "articles" that appear to be bulleted lists without the bullets. Apparently that is also a hallmark of AI? You get a few paragraphs and then the whole article turns into a sequence of one sentence paragraphs. Drives me batty!
I have an MFA in poetry. I was a journalist. I made an adult-living running a content management company. Yet, I read the article you refer to and I did not notice the AI. Only after a fellow Substacker pointed out similar issues that you have pointed out did I notice. I was embarassed. The issues are obvious. But I think a big reason I did not notice was that I was interested in his subject matter. So I wasn't looking for those issues. My bad. Yes, Embarassed! And Maybe I'm getting old and worn out with critiquing every aspect of my life, but I am forgiving of people who don't consider writing to be a strong part of who they are and rely on tech for help. Sure, the writer relied on too much AI. He didn't make it into his own voice. But he was honest about his use of AI. And the tool allowed him to express a concept that was important to him. I can forgive his use of AI for those reasons.
We each have our lines in the sands of technology: presets, anyone? auto-focus and image stabilization, y'all? Auto-spell, children? I say, choose your writing and reading art-creating comfort zone on a case-by-case basis and pick your battles.
Oh, the problem is everywhere, and that guy has a huge following, and everything is AI-generated, the same sentences, essay after essay with a few different words stuck in. My worry is that people who can write will not be able to get jobs, they will go to people who can’t organize a thought, but use Ai to send out their resume.
Like I said, I'm tired of hearing myself critique everything. I am moving to a view of the world in which I only open my mouth to that which i can control, and leave the rest to whatever definition of "progress" makes for IPOs and headlines. i'll look closer at his history of posts and will likely agree with you, since so many others I follow echo your sentiment. In the meantime I'll keep my head down and write the old fashioned way, and let AI generated content cannibalize its own feces. If Substack becomes overwhelmed with robotic thinking and image producing... Well, I'll figure out that step when it's time. The world needs warriors to keep us from straying too far. So thank you. Sincerely. I used to think I was one of them. Not no more.
Editing your own work usually means getting rid of 20-30 percent of what you wrote. Read it out loud while editing, print it, read it, and mark it up. Some AI’s can take your typed original along with your handwritten notes and make corrections for you that you handwrote. I have people I work with do this. You can also ask AI for feedback and then make the corrections yourself. Ask AI if it could ask you 5 questions that, if answered, would make your piece better. Regretfully, many of the “writers” we have here do not make corrections themselves and use AI to do the job for them rather than using it to help them make the changes themselves. Keep writing.
After a good night’s sleep, I have come to your side on this. Reading what seems to be an entire history of AI-engineered prose nearly broke me.
You might be astonished by the myriad Media I’ll peruse & filter through daily as a ‘hunter & gatherer’ - acolyte of Marshall Mcluhan - & ‘generalist
- Flood The Zone With Shit impacts far more than ‘just politics’ - I hired & was hired - over the years within those Media. Translators, Drivers, Makeup Artists, Musicians, Illustrators, Electricians, Programmers, Staging, Production Assistants, Animators, Technicians, Guides, Security, Editors, Wardrobe, Riggers, Typesetters.. & often renting the specialized Gear or / Services required by those experts - for just one day, an hour, even weeks or 15 seconds .. not to mention Talent / Stunt Actors
Those People ? AI is wiping them out daily - We cannot even ‘see them or hear them’ through the Shit Flooding into Social Media 24/7
To err is human. The idea that AI doesn’t err is false. It’s getting better and will continue to do so but I’d feel as though I was cheating to use it in my writing. I know my punctuation isn’t always correct and I did horribly in grammar in school. In the years that followed I found that I enjoyed writing. I’m hoping to improve in my skills but I want it to be all me, even if it takes longer to get my thoughts into words. Appreciate your thoughts here on the subject.
I often include a bibliography of my readings. I supply the details and AI is asked to put it into a specific format. It will remove a book name and replace it, tell me a quate is from another book when I tell it I have the book on my desk in front of me and know who wrote it and when, but didn’t want to write it up in a certain format it says, “Oh, good you caught that. You're right. I asked it to tell me the date that Charle Kirk was shot it told me he wasn’t shot, that I was misremembering, and that I must have some anxiety and should talk to someone. I said I know what I read and was told that people who are upset sometimes make mistakes and misunderstand things. I loaded a copy of the NYT’s for it to read and it said, “Sorry, I hadn’t updated that information.” I asked it what would happen if a 14 year old were told they had basic facts wrong and were emotionally confused if offered no comment.
Ooof. I don’t think I could’ve gotten through that person’s article with these examples you’ve pointed out here. I’ve never used AI myself so it was interesting to see how poorly it edits a text. Yeesh.