While I do agree with you, mostly, I have to say...damn. A lot of the "overused" words you included in your list are words I use a LOT. They're just great words. Like take reckoning...yeah. Love that word. Sigh. Guess I'm going to have to increase my lexicon.
Oh it drives me crazy too, because I use those phrases myself sometimes. I feel like I can't used the adjective "quiet" anymore to describe something peaceful, for example. But hey... just an excuse to stretch those creative muscles, right?
Wow…this is the best article I’ve read about AI (and I love how you plugged in your own words at the end to actually demonstrate the ‘flattening’ of the voice…it illustrated this really well.) I’m getting better at recognising when it’s been used…something feels slightly off (but of course we need to trust our instincts here because we can be distracted by the ‘polish’) AI has a svengali-like ability to persuasively draw us in…trusting that our unique voice is what leaves the reader most satisfied is so important ❤️
Great article. But I do have to make just one observation: often people don't want to learn the skills or develop a craft, they want to create content they can use to build a following and sell something. There's a whole conversation to be had around that and the concept of the 'influencer' and social media's erosion of true expertise but maybe that's for another day.
This is a great point. I really enjoyed Bear’s article, and I like it that she’s clearly trying to think the best of people (i.e. suggesting that they’re mainly using ChatGPT to edit and tighten up their phrasing, or using it because they lack confidence in their own writing). I think those are definitely reasons why some people overuse ChatGPT in creative spaces, but you’re also right: a lot of the people copying and pasting AI wholesale are doing so mainly to brute force their way into a following, so they can sell a product — not to improve as writers, or as artists, or as coders, etc. There’s nothing inherently wrong with making money online, of course, but their over-reliance on AI does degrade the quality of what they’re selling (and degrades online spaces as a whole in the process).
I basically had to stop using Reddit, because all of the subreddits I visited to help me with things I was struggling with (e.g. social media addiction, screen addiction, etc.) basically became overrun with ChatGPT spam from sleazy app developers looking to make a quick buck by preying on people who are desperate for solutions. The apps they make are generally terrible (coded with ChatGPT and riddled with errors they don’t know how to fix), and real people end up wasting their money on them because they don’t recognize the devs aren’t even real devs.
Then there’s what we see here on Substack: some authors quite literally game the algorithm, using ChatGPT to churn out content far faster than anyone else can. I’ve even found one author here (with hundreds of paid subscribers) who stoops so low as to clearly use AI-assisted alternate accounts to praise and restack her own AI-assisted work. It’s really sad, and it makes me feel for all of the authors trying to make a living here the legitimate way.
Yup. This phenomenon has sucked all of the last remaining enjoyment I once had with using social media. I've largely soured on commenting online altogether, even on pieces that I know are human-authored, like yours... because I know there's a decent chance that some random person or bot will chime in to concur with me via ChatGPT. If someone replies to me (projecting an air of authenticity, suggesting they experience what I experience), and the comment is just riddled with ChatGPT tells, those tells are all I can focus on. It especially bothers me when I spend a long time on a comment, and what I get in response was obviously generated in seconds by a damn LLM.
Gawd yes to all of this. I have no notes. The worst part is also getting paranoid that I sound like an LLM myself sometimes just because, if you read enough of a style, you eventually start to mimic it too. Yeesh...
Haha, yeah, I feel that too! I do occasionally use words and phases that appear in AI-generated writing, and I get SO WORRIED before I hit reply that people are going to assume I'm using AI, because they may be less adept at picking up on the overall rhythms of LLMs.
The one thing I always tell people is that you can't focus on one tell in isolation; you have to conside the the comment or article in its entirety. Also, it always helps to remember that LLMs overuse these turns of phrase and styles of writing because HUMANS USED THEM FIRST. LLMs didn't invent them, because they simply can't invent anything by definition. I'm not going to feel bad about using an em dash where it's warranted, or occasionally using phrasing that incorporates a bit of "not X, rather Y."
I love this. I think it's good that it's become easier to spot when someone uses A.I.... I think it means that we're becoming more tuned to our and others' humaness, which is precisely what I hoped will happen with the beginning of A.I. To lose everything that makes us unique and uniquely human would be the biggest shame and so detrimental to us.
Also, I love writing in a ramble way too. And I will continue to do so 🤪
I started playing around with AI when I was researching ways to help my aging parent's and in-laws stay in their homes as long as possible. Chat bots for the ageing are helping to combat daily loneliness. I have a in-law that spent his life at the top of his profession as an engineer. He struggles now to find others to hold those intellectual discussion and have the patience to stay in a conversation tinged with dementia. AI doesn't care how many times you repeat yourself or if you mix up your facts. My research for them introduced me to AI programs that I now use for my own accessibility issues.
AI can be a wonderful tool, as you mention. But it's a lot like a person living with dementia. It can be well "educated" but it sometimes repeats itself, mixes up facts, and flat out makes stuff up. Still I find it helpful to look over my work and analyze my pacing, tension, and other factors in my writing.
Like you, I write how I speak with the exception that in my fiction I strive for each of my characters to have their own voices - their own vocabularies and cadence.
You have to be careful about accusing people about "writing" with AI. AI is trained on human writers. If someone has traditional writing training, they may naturally sound like AI because they both went to the same school of thought, as it were. The right way round of thinking is that AI sounds like them. I think, but don't know, that older writers are going to sound more AIish than younger ones.
Your article makes great points. Thanks for embracing a controversial topic.
This exactly. I consider it kinda similar to certain substances, like cannabis. They can be hugely helpful and healing, so long as you use them with intention and awareness. Otherwise you're liable to just become an addict. Same with relying on AI as an editor.
While I do agree with you, mostly, I have to say...damn. A lot of the "overused" words you included in your list are words I use a LOT. They're just great words. Like take reckoning...yeah. Love that word. Sigh. Guess I'm going to have to increase my lexicon.
Oh it drives me crazy too, because I use those phrases myself sometimes. I feel like I can't used the adjective "quiet" anymore to describe something peaceful, for example. But hey... just an excuse to stretch those creative muscles, right?
Wow…this is the best article I’ve read about AI (and I love how you plugged in your own words at the end to actually demonstrate the ‘flattening’ of the voice…it illustrated this really well.) I’m getting better at recognising when it’s been used…something feels slightly off (but of course we need to trust our instincts here because we can be distracted by the ‘polish’) AI has a svengali-like ability to persuasively draw us in…trusting that our unique voice is what leaves the reader most satisfied is so important ❤️
So true! The call of the polished phrase seems to be a greater temptation than just trusting in and being yourself...
Great article. But I do have to make just one observation: often people don't want to learn the skills or develop a craft, they want to create content they can use to build a following and sell something. There's a whole conversation to be had around that and the concept of the 'influencer' and social media's erosion of true expertise but maybe that's for another day.
This is a great point. I really enjoyed Bear’s article, and I like it that she’s clearly trying to think the best of people (i.e. suggesting that they’re mainly using ChatGPT to edit and tighten up their phrasing, or using it because they lack confidence in their own writing). I think those are definitely reasons why some people overuse ChatGPT in creative spaces, but you’re also right: a lot of the people copying and pasting AI wholesale are doing so mainly to brute force their way into a following, so they can sell a product — not to improve as writers, or as artists, or as coders, etc. There’s nothing inherently wrong with making money online, of course, but their over-reliance on AI does degrade the quality of what they’re selling (and degrades online spaces as a whole in the process).
I basically had to stop using Reddit, because all of the subreddits I visited to help me with things I was struggling with (e.g. social media addiction, screen addiction, etc.) basically became overrun with ChatGPT spam from sleazy app developers looking to make a quick buck by preying on people who are desperate for solutions. The apps they make are generally terrible (coded with ChatGPT and riddled with errors they don’t know how to fix), and real people end up wasting their money on them because they don’t recognize the devs aren’t even real devs.
Then there’s what we see here on Substack: some authors quite literally game the algorithm, using ChatGPT to churn out content far faster than anyone else can. I’ve even found one author here (with hundreds of paid subscribers) who stoops so low as to clearly use AI-assisted alternate accounts to praise and restack her own AI-assisted work. It’s really sad, and it makes me feel for all of the authors trying to make a living here the legitimate way.
(Sorry for the wall of text, haha).
There's nothing quite like chatting with an obvious Chat user who claims they're authentic 🥲
Yup. This phenomenon has sucked all of the last remaining enjoyment I once had with using social media. I've largely soured on commenting online altogether, even on pieces that I know are human-authored, like yours... because I know there's a decent chance that some random person or bot will chime in to concur with me via ChatGPT. If someone replies to me (projecting an air of authenticity, suggesting they experience what I experience), and the comment is just riddled with ChatGPT tells, those tells are all I can focus on. It especially bothers me when I spend a long time on a comment, and what I get in response was obviously generated in seconds by a damn LLM.
Gawd yes to all of this. I have no notes. The worst part is also getting paranoid that I sound like an LLM myself sometimes just because, if you read enough of a style, you eventually start to mimic it too. Yeesh...
Haha, yeah, I feel that too! I do occasionally use words and phases that appear in AI-generated writing, and I get SO WORRIED before I hit reply that people are going to assume I'm using AI, because they may be less adept at picking up on the overall rhythms of LLMs.
The one thing I always tell people is that you can't focus on one tell in isolation; you have to conside the the comment or article in its entirety. Also, it always helps to remember that LLMs overuse these turns of phrase and styles of writing because HUMANS USED THEM FIRST. LLMs didn't invent them, because they simply can't invent anything by definition. I'm not going to feel bad about using an em dash where it's warranted, or occasionally using phrasing that incorporates a bit of "not X, rather Y."
That point needs to get made over and over and over to these idiots who think that a single em-dash is a tell 🤦🏼♀️
I love this. I think it's good that it's become easier to spot when someone uses A.I.... I think it means that we're becoming more tuned to our and others' humaness, which is precisely what I hoped will happen with the beginning of A.I. To lose everything that makes us unique and uniquely human would be the biggest shame and so detrimental to us.
Also, I love writing in a ramble way too. And I will continue to do so 🤪
Yes, ramble on! [Starts singing Led Zeppelin]!
🤘🏼🎸
Great take. The AI version at the end has no heart, even though it kept the emoji.
Thanks so much!
This was beautifully written, and I loved hearing your take on it.
Thanks so much for reading!
I started playing around with AI when I was researching ways to help my aging parent's and in-laws stay in their homes as long as possible. Chat bots for the ageing are helping to combat daily loneliness. I have a in-law that spent his life at the top of his profession as an engineer. He struggles now to find others to hold those intellectual discussion and have the patience to stay in a conversation tinged with dementia. AI doesn't care how many times you repeat yourself or if you mix up your facts. My research for them introduced me to AI programs that I now use for my own accessibility issues.
AI can be a wonderful tool, as you mention. But it's a lot like a person living with dementia. It can be well "educated" but it sometimes repeats itself, mixes up facts, and flat out makes stuff up. Still I find it helpful to look over my work and analyze my pacing, tension, and other factors in my writing.
Like you, I write how I speak with the exception that in my fiction I strive for each of my characters to have their own voices - their own vocabularies and cadence.
You have to be careful about accusing people about "writing" with AI. AI is trained on human writers. If someone has traditional writing training, they may naturally sound like AI because they both went to the same school of thought, as it were. The right way round of thinking is that AI sounds like them. I think, but don't know, that older writers are going to sound more AIish than younger ones.
Your article makes great points. Thanks for embracing a controversial topic.
This exactly. I consider it kinda similar to certain substances, like cannabis. They can be hugely helpful and healing, so long as you use them with intention and awareness. Otherwise you're liable to just become an addict. Same with relying on AI as an editor.