or, How Drinking Cappuccino at a Little Café in Paris overlooking the Eiffel Tower... - A few ideas on how to avoid unjust accusations of AI in your writing.
A bit off topic, but...I was talking to my brother (who teaches people how to use AI to work more productively) about the ethics of using AI to write. He feels some guilt around using it to write his presentations - like it's stealing from the writers whose work has been scraped - although that won't stop him from using it.
My take (and maybe a bit more on-topic) is that my writing is a personal connection to my readers, I've worked hard to find my voice, and even if I used AI to replicate my voice, I would view it as a betrayal to my relationship to my readers.
Maybe this is just a roundabout way of saying, "develop your voice and trust it." And of course, your voice when working for clients will be different (and more AI-able) than in your own work.
"Develop your voice and trust it" - I agree completely!
And yeah, people are going to use AI. It's unavoidable. As someone pointed out to me recently, even if you don't think you're using AI, something you're using is using AI...
All we can do is learn how to use it ethically and responsibly. Like all tools, I suppose!
I'm not going to feed my stuff to AI so it can tell me if AI wrote it or me. That seems self defeating... also, now the damn thing has my writing to chew on. Won't do it! I already tend to use too many fragments, so I guess I'm OK. Screw the thing about em-dashes. I finally figured how to get them on my Mac and I'm not going to unlearn that, it was too painful to begin with. As to typos... aargh... typos give me heartburn. I won't let them slip by (to the best of my knowledge). Nobody will say I'm sloppy even if that makes me machine-adjacent. In summary, give me a Luddite badge: I'll keep the robots away with a pitchfork.
Yes -- I've gotten an earful about typos! lol If you knew me in real life, you wouldn't believe I wrote that... As I mentioned to someone else I was talking to, these are just things you *can* do. I'm not promoting them as things you *should* do!
My comment has always been that if my writing doesn't pass the AI detector, it's simply proof that they used my writing to train AI.
Hmm. Sounds like you described the way AI pushes human writing around by sneaking in the back door. "AI doesn't do this, so do it." Misspelling!? Neologisms!? I see and understand the rationale, but I also disagree ... without knowing or offering a better alternative.
Looking at the trajectory of LLMs since, say ChatGPT in November 2022 to now, you can see "improvement" in AI mimicry of the human voice. The models have incorporated styles, too. The area of human occupation, so to speak--where AI can't enter (yet)--has narrowed. And to prove our humanity, we need to emphasize our foibles? Our quirks? Some forced nuance?
I wonder if we should all go punk. Righteous indignation sounds about right. The Luddites used hammers. Maybe they we on to something.
I'm quite suspicious about using LLMs, having just toyed around with them. I do see their use in transcription, though. Gemini can cough up a reasonable transcription of a recording, with errors of course, in a couple minutes. That's a real timesaver!
Yes, it makes me feel icky, too. And in all honesty, I'm not worried about my writing being confused with AI. I've never used an AI detector because (a) I know I wrote it and (b) I know AI detectors are deeply flawed.
However, I do know that many writers struggle with this -- especially newer ones who don't yet have a track record to back up their work. The reality is that writers are getting called out, often unjustly, for their writing. I've personally talked to at least one writer about this issue, and heard many other stories.
As to AI/LLMs in general -- I think you and I have had this discussion before: I have seen the power and potential of AI for over a decade now. I'm not suspicious of the technology per se, but the people behind the technology make me nervous. The system too easily manipulated (never mind the proven biases baked into the code), and the data used to train these models is often stolen.
(Cue righteous indignation!)
The reality is, AI is here to stay, regardless of how we feel about it. I'm not exactly sure myself how to navigate this new landscape. But differentiating human writing from AI -- even through icky means, maybe -- is at least a start!
Maybe if we all stop talking about AI it will just get bored and go away?
That's a theory I'd be willing to test. 😂
A bit off topic, but...I was talking to my brother (who teaches people how to use AI to work more productively) about the ethics of using AI to write. He feels some guilt around using it to write his presentations - like it's stealing from the writers whose work has been scraped - although that won't stop him from using it.
My take (and maybe a bit more on-topic) is that my writing is a personal connection to my readers, I've worked hard to find my voice, and even if I used AI to replicate my voice, I would view it as a betrayal to my relationship to my readers.
Maybe this is just a roundabout way of saying, "develop your voice and trust it." And of course, your voice when working for clients will be different (and more AI-able) than in your own work.
"Develop your voice and trust it" - I agree completely!
And yeah, people are going to use AI. It's unavoidable. As someone pointed out to me recently, even if you don't think you're using AI, something you're using is using AI...
All we can do is learn how to use it ethically and responsibly. Like all tools, I suppose!
You're not peddling that old panacea of common sense, are you?!?!
I'm not going to feed my stuff to AI so it can tell me if AI wrote it or me. That seems self defeating... also, now the damn thing has my writing to chew on. Won't do it! I already tend to use too many fragments, so I guess I'm OK. Screw the thing about em-dashes. I finally figured how to get them on my Mac and I'm not going to unlearn that, it was too painful to begin with. As to typos... aargh... typos give me heartburn. I won't let them slip by (to the best of my knowledge). Nobody will say I'm sloppy even if that makes me machine-adjacent. In summary, give me a Luddite badge: I'll keep the robots away with a pitchfork.
Yes -- I've gotten an earful about typos! lol If you knew me in real life, you wouldn't believe I wrote that... As I mentioned to someone else I was talking to, these are just things you *can* do. I'm not promoting them as things you *should* do!
My comment has always been that if my writing doesn't pass the AI detector, it's simply proof that they used my writing to train AI.
Hmm. Sounds like you described the way AI pushes human writing around by sneaking in the back door. "AI doesn't do this, so do it." Misspelling!? Neologisms!? I see and understand the rationale, but I also disagree ... without knowing or offering a better alternative.
Looking at the trajectory of LLMs since, say ChatGPT in November 2022 to now, you can see "improvement" in AI mimicry of the human voice. The models have incorporated styles, too. The area of human occupation, so to speak--where AI can't enter (yet)--has narrowed. And to prove our humanity, we need to emphasize our foibles? Our quirks? Some forced nuance?
I wonder if we should all go punk. Righteous indignation sounds about right. The Luddites used hammers. Maybe they we on to something.
I'm quite suspicious about using LLMs, having just toyed around with them. I do see their use in transcription, though. Gemini can cough up a reasonable transcription of a recording, with errors of course, in a couple minutes. That's a real timesaver!
Yes, it makes me feel icky, too. And in all honesty, I'm not worried about my writing being confused with AI. I've never used an AI detector because (a) I know I wrote it and (b) I know AI detectors are deeply flawed.
However, I do know that many writers struggle with this -- especially newer ones who don't yet have a track record to back up their work. The reality is that writers are getting called out, often unjustly, for their writing. I've personally talked to at least one writer about this issue, and heard many other stories.
As to AI/LLMs in general -- I think you and I have had this discussion before: I have seen the power and potential of AI for over a decade now. I'm not suspicious of the technology per se, but the people behind the technology make me nervous. The system too easily manipulated (never mind the proven biases baked into the code), and the data used to train these models is often stolen.
(Cue righteous indignation!)
The reality is, AI is here to stay, regardless of how we feel about it. I'm not exactly sure myself how to navigate this new landscape. But differentiating human writing from AI -- even through icky means, maybe -- is at least a start!