Substack added AI detection. But the scan button isn't the threat. What readers were already sensing, plus the 4-step workflow that keeps your voice yours.
I write my articles using ChatGPT to take my thoughts, experiences, and ideas to create full length articles. All ideas are authentic, AI is my ghost writer.
There’s absolutely nothing wrong with that. I had AI tweak a few things in my last article and if you look up to see if it was AI written, it says about 15%. Big deal. So I needed a little help.
Been watching, and adjusting accordingly. I don't shy away from the fact that I use AI as a creative, but I've always been a writer, and you can't take that from me. There's a huge difference between directing AI, and copy n pasting whatever, we don't do that round here...✍️🏾🦁
Nicky, “automate the production, never the perspective” is where your piece struck a chord with me.
I recently wrote about Pangram after running my own experiment. My original fiction was labeled 100% AI, while an AI rewrite of the same work was labeled 100% human.
The incorrect scores didn’t linger in my mind. Instead, I couldn’t shake the transformation that occurred before I even scanned the writing. The AI version maintained the facts but removed any sense of judgment. Characters turned into mere explanations of themselves, losing their depth. A city I had dedicated months to building suddenly felt like a description rather than a vibrant place filled with life and experience.
That made me think the deeper question is not simply whether AI touched the work. It is whether the writer still owns the decisions inside it.
Perspective, judgment, omission, memory, specificity, knowing what not to explain: those are difficult things for a detector to measure.
So I’m with you on protecting perspective. I would only add that transparency tools should never become substitutes for reading the work itself.
‘Automate the production, never the perspective’ really stood out to me. I use AI as a writing partner, but I still want the ideas, experiences, opinions, and final judgment to be mine. The goal for me isn’t to avoid AI; it’s to make sure I don’t disappear from my own writing.
Thank you for this! I also think it's important to note WHY this is happening now. This is NOT to protect us - the little people. It is to protect the LLMs. The AI models are scanning human written text to learn from. If all of the text is AI you will eventually have model collapse. AI cannot be trained on AI - it has to be trained on human writing. So the invisible water mark lets the AI know not to scan that particular text. We are unwittingly feeding the models with our writing. The tech bros are winning again. I guess we are winning by our frontal lobes not turning into mush. But all in all -this is about protecting AI - not humans.
I totally agree with you. Even as I write, it automatically attempts to anticipate my next words. It is learning as it goes. Even our phones do autocorrect. Nice!!!
I use AI almost every day for my online content creation, and I have no problem saying so.
English is not my strongest point, and without AI, some of the ideas in my head may not come out as clearly as I want them to.
But I don't simply ask AI, “Write me an article about this,” copy it and publish it.
I talk to AI.
I share my ideas, thoughts, experiences, wishes and sometimes even dreams about something I want to write about. I question it. I disagree with it. I change things. We go back and forth until what comes out actually sounds like something I would say — just without my broken English or grammar problems. 😄
So if an AI detector tells me that my article is highly AI-written, I am honestly not too bothered.
The words may have been helped by AI, but the **thinking, direction and final decision are still mine**.
For me, AI is not about pretending I can write perfect English.
It is about finally having a tool that allows the ideas inside my head to come out in the way I have always wanted to communicate them.
And as long as the finished article still speaks at **my level, in my way, and represents what I genuinely think**, I am comfortable putting my name on it. 😊
Nicky, this raises an important distinction for me: a thought is still a thought, irrespective of the technology used to help express it.
A startup founder may have mentors who help refine the strategy or presentation, but the original vision can still belong to the founder. I see AI similarly—as a tool that can help organize, challenge and refine expression, while the underlying idea, lived perspective, intent and responsibility remain human.
That is why I’ve started adding a simple Creator’s Note to my own reflections: the central ideas, perspective, experience and intent originate with me; AI is used only to help organize and refine the wording.
Perhaps the deeper question is not simply “Was AI used?” but “Where did the thought originate, and who stands behind it?”
Yep, I noticed that this morning when I posted a Substack article. I don’t trust AI, so I will totally do a brain dump, and talk to it every step of the way until I’m satisfied. It’s been very helpful, but I heard some people are saying, written, in collaboration with AI, and I also noticed that there was a turn off feature before you hit the publish button. Still learning about that.
I’m just starting with Substack and have dyslexia, easily misspell words, even the sentence structure comes out sounding backwards. I’ve been using AI to check spelling, typing errors, grammar and flow of the text. Often I have read the suggestions and I’m like well said instead of my jumbled run on sentences. Is it best to state written with AI assistance or maybe I should give up entirely?Which would be a shame as I feel like I have important helpful experiences to share.
I have dyslexia too and have the same issue. Some people don't like my weird sentence structure. On the other hand, I am weird in many ways and want that to come through in my writing. I want the people that can't see through odd sentence structure to fall away, and the people who can see through that to connect with me. So, to me it's an advantage to have, that AI would take away from me.
I do work on my writing and have learned to put a full stop where my comma is and start a new sentence. Where possible. That habit has made me slow down when writing and organize my thoughts better. It's also given me more confidence.
I don’t see it as a problem. I see AI as an amplifier. If your content is rubbish, AI will amplify your “rubbishness"!!! haha There’s been ghostwriters online since ages too, up to everyone to use AI the smart way to remove the AI slop!
Honestly, I feel like AI was fine when companies were the only beneficiaries of it. They’ve been using it as a process of elimination for years now when it comes to jobs. But now that the general public is using it to “get jobs”, and other life needs, they have a problem with it. It’s like whenever we figure out how to simplify our lives, they find a way to make it harder.
I’ve heard how AI has helped people advocate for themselves in both medical and legal ways, so now it’s putting pressure on those folks.
The truth is they weren’t expecting us to get good at it and use it to our benefit.
LOL, go to substack support page, click on 'Help' > contact support. You'll get this message, "The Substack chatbot is here to answer your questions!"! And I didn't have to use AI detection to know they are using AI.
Is AI flagging AI or work that is already online? I read this article where its stated that "this wont affect preexisting material," but will it flag preexisting material being reprinted, if that material is your own work? This is where TURNITIN fails. I have numerous articles in the cloud, a book, and many social media posts on theological subjects and wonder if AI detects AI or anything that already exists online?
I use AI as an editor. I feed it my first draft and it helps with polish. It does not write for me or add any big ideas, but it does correct my grammar and clear spelling mistakes.
I write my articles using ChatGPT to take my thoughts, experiences, and ideas to create full length articles. All ideas are authentic, AI is my ghost writer.
this is exactly my thought and process.
Agreed with what @Kiara Ameeka had written.
There’s absolutely nothing wrong with that. I had AI tweak a few things in my last article and if you look up to see if it was AI written, it says about 15%. Big deal. So I needed a little help.
Right. We shouldn’t hide it. AI is a tool just like Microsoft Word is a tool.
Agree!
Been watching, and adjusting accordingly. I don't shy away from the fact that I use AI as a creative, but I've always been a writer, and you can't take that from me. There's a huge difference between directing AI, and copy n pasting whatever, we don't do that round here...✍️🏾🦁
Preach!!
Facts
Nicky, “automate the production, never the perspective” is where your piece struck a chord with me.
I recently wrote about Pangram after running my own experiment. My original fiction was labeled 100% AI, while an AI rewrite of the same work was labeled 100% human.
The incorrect scores didn’t linger in my mind. Instead, I couldn’t shake the transformation that occurred before I even scanned the writing. The AI version maintained the facts but removed any sense of judgment. Characters turned into mere explanations of themselves, losing their depth. A city I had dedicated months to building suddenly felt like a description rather than a vibrant place filled with life and experience.
That made me think the deeper question is not simply whether AI touched the work. It is whether the writer still owns the decisions inside it.
Perspective, judgment, omission, memory, specificity, knowing what not to explain: those are difficult things for a detector to measure.
So I’m with you on protecting perspective. I would only add that transparency tools should never become substitutes for reading the work itself.
‘Automate the production, never the perspective’ really stood out to me. I use AI as a writing partner, but I still want the ideas, experiences, opinions, and final judgment to be mine. The goal for me isn’t to avoid AI; it’s to make sure I don’t disappear from my own writing.
Thank you for this! I also think it's important to note WHY this is happening now. This is NOT to protect us - the little people. It is to protect the LLMs. The AI models are scanning human written text to learn from. If all of the text is AI you will eventually have model collapse. AI cannot be trained on AI - it has to be trained on human writing. So the invisible water mark lets the AI know not to scan that particular text. We are unwittingly feeding the models with our writing. The tech bros are winning again. I guess we are winning by our frontal lobes not turning into mush. But all in all -this is about protecting AI - not humans.
@Lisa Woods; well said; that’s what I was thinking, they only want humanlike text so we can train it;
they are playing on our emotions.
I totally agree with you. Even as I write, it automatically attempts to anticipate my next words. It is learning as it goes. Even our phones do autocorrect. Nice!!!
I use AI almost every day for my online content creation, and I have no problem saying so.
English is not my strongest point, and without AI, some of the ideas in my head may not come out as clearly as I want them to.
But I don't simply ask AI, “Write me an article about this,” copy it and publish it.
I talk to AI.
I share my ideas, thoughts, experiences, wishes and sometimes even dreams about something I want to write about. I question it. I disagree with it. I change things. We go back and forth until what comes out actually sounds like something I would say — just without my broken English or grammar problems. 😄
So if an AI detector tells me that my article is highly AI-written, I am honestly not too bothered.
The words may have been helped by AI, but the **thinking, direction and final decision are still mine**.
For me, AI is not about pretending I can write perfect English.
It is about finally having a tool that allows the ideas inside my head to come out in the way I have always wanted to communicate them.
And as long as the finished article still speaks at **my level, in my way, and represents what I genuinely think**, I am comfortable putting my name on it. 😊
This is exactly how I use my AI 😅.
I think a lot and most times it gets really messy, so I use AI to organize my thoughts and put them in a normal essay format.
Yeah. At least it stop me from procrastinating because I simply can write well with my “poor English grammar”. 😃
Nicky, this raises an important distinction for me: a thought is still a thought, irrespective of the technology used to help express it.
A startup founder may have mentors who help refine the strategy or presentation, but the original vision can still belong to the founder. I see AI similarly—as a tool that can help organize, challenge and refine expression, while the underlying idea, lived perspective, intent and responsibility remain human.
That is why I’ve started adding a simple Creator’s Note to my own reflections: the central ideas, perspective, experience and intent originate with me; AI is used only to help organize and refine the wording.
Perhaps the deeper question is not simply “Was AI used?” but “Where did the thought originate, and who stands behind it?”
Yep, I noticed that this morning when I posted a Substack article. I don’t trust AI, so I will totally do a brain dump, and talk to it every step of the way until I’m satisfied. It’s been very helpful, but I heard some people are saying, written, in collaboration with AI, and I also noticed that there was a turn off feature before you hit the publish button. Still learning about that.
Yeah I’m big on the collaboration piece with AI. I’m with you i don’t trust it either.
Update, just saw in settings that you can add an AI statement.
I’m just starting with Substack and have dyslexia, easily misspell words, even the sentence structure comes out sounding backwards. I’ve been using AI to check spelling, typing errors, grammar and flow of the text. Often I have read the suggestions and I’m like well said instead of my jumbled run on sentences. Is it best to state written with AI assistance or maybe I should give up entirely?Which would be a shame as I feel like I have important helpful experiences to share.
I have dyslexia too and have the same issue. Some people don't like my weird sentence structure. On the other hand, I am weird in many ways and want that to come through in my writing. I want the people that can't see through odd sentence structure to fall away, and the people who can see through that to connect with me. So, to me it's an advantage to have, that AI would take away from me.
I do work on my writing and have learned to put a full stop where my comma is and start a new sentence. Where possible. That habit has made me slow down when writing and organize my thoughts better. It's also given me more confidence.
Just a thought!
This is something to consider. I do feel that we can turn what might be considered a hindrance into a strength. Thanks for that perspective.
I don’t see it as a problem. I see AI as an amplifier. If your content is rubbish, AI will amplify your “rubbishness"!!! haha There’s been ghostwriters online since ages too, up to everyone to use AI the smart way to remove the AI slop!
Honestly, I feel like AI was fine when companies were the only beneficiaries of it. They’ve been using it as a process of elimination for years now when it comes to jobs. But now that the general public is using it to “get jobs”, and other life needs, they have a problem with it. It’s like whenever we figure out how to simplify our lives, they find a way to make it harder.
I’ve heard how AI has helped people advocate for themselves in both medical and legal ways, so now it’s putting pressure on those folks.
The truth is they weren’t expecting us to get good at it and use it to our benefit.
You can also disable it 👌🏼
I use it to dictate my articles.
LOL, go to substack support page, click on 'Help' > contact support. You'll get this message, "The Substack chatbot is here to answer your questions!"! And I didn't have to use AI detection to know they are using AI.
Is AI flagging AI or work that is already online? I read this article where its stated that "this wont affect preexisting material," but will it flag preexisting material being reprinted, if that material is your own work? This is where TURNITIN fails. I have numerous articles in the cloud, a book, and many social media posts on theological subjects and wonder if AI detects AI or anything that already exists online?
I use AI as an editor. I feed it my first draft and it helps with polish. It does not write for me or add any big ideas, but it does correct my grammar and clear spelling mistakes.