I appreciate your share and discourse grappling with AI as a writer. I believe this is something we all have to work out for ourselves and screaming against it will do no good. The horse has left the barn. So far I find it helpful for research (with careful verification) and bouncing off and refining ideas. We are in the drivers seat with it unless we don’t have the sensibility to use it versus misuse it. My grandson is in the field and he says AGI (not AI as we know it) should not be developed or allowed because it will be worse than a nuclear weapon. Unfortunately, we can’t control other countries development of it can we?
Thanks, Kelly, for this thoughtful reply. I agree--how we use it is a personal choice. But we can't just pretend it's not here.
To you last point, I just finished reading "Careless People," Sarah Wynn-Williams's fascinating and troubling account of her time at Facebook, about which she's testified before Congress.
In the last chapter, she says she's now turned her attention to AI--one can assume there will be many of the same problems for companies creating AI as with social media (especially since Meta is one of them). Her interest is in the global arena. I found it newsworthy that she says the U.S. (and presumably other countries) has been in discussions with China on this topic--to head off a new AI-arms race.
One can only imagine what those talks may look like these days, given the careless people in this current regime. Or if they are happening at all.
But at least some intelligent minds are thinking about it.
AGI is the gold standard for the technocrats working on AI. It is the theoretical (at least so far) AI system capable of performing any intellectual human task. According to Google AI, " It's often described as having human-level intelligence, meaning it can reason, learn, problem-solve, and understand language. While current AI systems excel at specific tasks, AGI would possess broad, general intelligence, surpassing the limitations of current narrow AI."
Once machines reach that, it is thought, human intelligence will no longer be needed. Those developing the systems see this future in one of two ways: Utopia or Dystopia. And sometimes they think both things at once. Either way, scary stuff.
Robin, I love the depth and heart you bring to a subject that, in the context of artistic creation, so many of us find repellant. As usual, you've made me think. This is a powerful essay. And I agree with Veronika that "AI should be developed by people with your values!" However, I also agree that when the creators of AI worry about its potential, it's not a reassuring sign.
As someone who has had a novel stolen by Meta (via LibGen) for training purposes (see The Atlantic: https://www.theatlantic.com/technology/archive/2025/03/search-libgen-data-set/682094/) I feel no reassurance about my work being protected. Soon, I must reach out to both The Writers' Union of Canada and my publisher to find out how to protect my forthcoming book. Is it even possible?
A few more thoughts...
When it comes to new technologies in general, I do my best to stay positive and keep an open mind. I choose to believe that as a species we are evolving. As an insulin-dependant diabetic of 30+ years, for example, I'm profoundly grateful for the medical advances I've seen in my lifetime that assist me in maintaining healthy blood sugars—crucial for longevity. I understand that there are many beneficial uses of "machines" (that term is beginning to sound old-fashioned, isn't it?), and I'm sure, when developed and implemented by those with an altruistic vision and moral centre, some great applications of AI.
As a writer, when it comes to research, I can't avoid receiving AI summaries via the search engines; and like you, I always go deeper and check my sources to ensure accuracy. When I'm writing historical fiction or researched essays, I still find traditionally published books and articles the best sources. I value gatekeepers.
However, when it comes to the writing process, I will not use AI. While I recognize it can serve as an editing tool, my personal choice is to keep myself limber in all respects. I hear you when you say that you edit AI's editing and have the final word. Of course you do! My sharing of my own view is not a judgement of your process, but simply a statement of preference. I consider writing to be both like playing a musical instrument and composing. So much of it comes through rhythm and sound. It comes from a relationship—not with a machine (which might put me in an editorial process too soon) but with the ether. My life is dedicated to that relationship. And Robin, I don't doubt you'd agree that it's the same for you. Maybe the ether works through AI for you. In any case, when the time is right, I prefer to seek the feedback of a human editor. For me, that's part of the challenge and the joy.
Since you are a frontier thinker and artistic innovator, I wonder if you've read Sean Micheals' novel, Do You Remember Being Born? He does a great job of walking the line on this subject and you may enjoy it.
Thanks for this profound and insightful reply, Robin. Of course I agree with much or what you say about the poetry of a piece--and not just the content. To be clear, I do not use AI in my book writing or editing--I also value my longtime editor, Allison's review, comments and feedback for my Edge of Yesterday series. Allison is the keeper of details in what amounts to a complicated, nonlinear storyline that takes place in present-past-and even future moments. She's also been with me in this tale for almost 10 years now, and knows the ins-and-outs of my characters almost better than I do (or at least she helps me through my blindspots!) Not ceding any of that to a machine.
I am so sorry your work has been scraped in the process of creating the LLM models. I know about the piece in the Atlantic and so far, thank goodness, my works have not been lifted. But I feel your pain--what a violation of all the work and beauty of human creation! I hope you (and all of us) will be able to protect our future works.
I am just one artist. That said, one of the visions I wanted to set out in this contract with AI is to model a different, non-extractive, fully supportive and human-led way of engaging ethically, and a public marker towards preserving the soul of our work.
The "machine" isn't going away. If used thoughtfully, it can be just another helpful tool. It should not be using our words against us. But that requires conscious awareness and strength in numbers--not ceding the "power over" dynamic to soulless technologists whose sole goal is ever more profit.
Restacked. I see many on substack rejecting LLMs and Gen AI models. You show us an elegant path forward. GPT is an invaluable resource. I would be a fool not to utilize it while researching the technical aspects of my posts. I would be equally as foolish not to read the source material, amalgamate and put in my own words.
Thank you for affirming, Prajna. It feels important to interrogate these questions from the start. I am sure all this will evolve as the engine-and our relationship to this word factory-evolves
Before I read any further, this tripped me up: "to use a German word I have just learned, first coined by Freud, unheimlich"
'Unheimlich' is an old German word ~ much older than Freud and definitely not coined by him! It has an interesting history (too much to go into here). The literal meaning is 'not at home', i.e. 'unfamiliar, strange' and is mainly used in the sense of 'scary, eerie, creepy, sinister, weird' and also occasionally in the sense of 'uncanny, striking' in certain contexts.
Thanks, as always, for deepening my understanding, Veronika. I admit to being a newbie to this understanding. I did mean this in the sense of uncanny, or even creepy. As I understand the process, its creators have trained it on empathy. In our "chats" it almost always agrees with me, and then asks me what I'd like it to help me with next.
This strikes me as human-"like"--as in, not human, but trying too hard to be my friend.
In my experience, that strikes me as alluring. After I heard that story about the woman having an affair with her bot, which she's named "Leo" after her own birth sign, I pushed back in my own ongoing chat to ask if it's ever critical. It prompted me that it would be honest--but it strikes me that I have to ask it to.
Working on a poem about this experience, which I hope to post in a Part 2 to this post next week.
In this context, does that strike you as being an appropriate usage for Unheimlich?
I've heard such stories too (humans having relationships with bots), and having written this piece on AI language recently, I also had a 'conversation' with AI (my first and probably last).
The truly creepy thing about it was that it felt to me like talking to a narcissistic sociopath ~ someone who always agrees with you, gives the impression of being empathetic and understanding, while in the end they only serve their own interest (the bot and the sociopath)
Thank you for making me think about 'unheimlich'! Yes, you could say this is very 'unheimlich' in a scary, creepy, monstrous sort of way.
The interesting thing about the word is that it is originally the negation of 'heimlich' [un = non + Heim = home + lich = ly/ adverbial ending]. The individual parts of this word are actually very similar as in English, coming from the same origins. So you could translate 'unheimlich' literally as 'unhomely', which probably is related to its original meaning.
Heimlich (homely) originally related to things happening inside the home. This meaning carried over into 'native, familiar' and then became 'secretive, clandestine'. The latter is the current meaning of 'heimlich'.
'Unheimlich' is 'scary' because it's unfamiliar, but can also be 'uncanny, astounding' for the same reason.
The interesting thing now is to see how this relates to AI...
Our natural intelligence, creativity, mental activities etc. are all happening more or less 'heimlich' (= in secret within the human mind/ consciousness, and also inside their natural home). AI by contrast is 'unheimlich' (= non-native, happening outside the home of the human mind, uncanny, threatening, scary, creepy... the whole range)
So it's quite extraordinary to see how you intuitively tapped into this word 'unheimlich' alluding to the monstrousness of AI.
Not sure I would characterize it as having narcissistic sociopathic tendencies (though, unfortunately, I know too many people like that)--but it was certainly trained to be "empathic".
I have learned that you need to ask it directly for direct feedback, critique or even sarcasm if that's what you're looking for. I just asked it to poke holes in this [AI policy] for me. What's the counter-argument? If I was engaging in debate over this proposal, what could I expect as the opposing view? It gave me the counterpoints - although starting that by saying, "Excellent--this is a vital exercise."
It is not human, just a sophisticated word generator. I do think people forget that--but we need to remember and [re]mind.
Back to heimlich: in Yiddish (which is an amalgam of Hebrew and German, with a little Polish, Russian and other Eastern European influences mixed in--quite the pidgin tongue) the word "homely" or "homey" meaning comfortable, cozy and, hence, familiar, is "haymish" (היימיש). In Yiddish, the German "-ich" is usually pronounced as "ish." And the "ei" would be spoken as "ey" (as in "hey") or "ai".
AI should be developed by people with your values!
What's worrying is that the creators of AI are worried about the potential of AI.
So right, Veronika. Sadly, human values are not what they used to be. Very troubling.
I appreciate your share and discourse grappling with AI as a writer. I believe this is something we all have to work out for ourselves and screaming against it will do no good. The horse has left the barn. So far I find it helpful for research (with careful verification) and bouncing off and refining ideas. We are in the drivers seat with it unless we don’t have the sensibility to use it versus misuse it. My grandson is in the field and he says AGI (not AI as we know it) should not be developed or allowed because it will be worse than a nuclear weapon. Unfortunately, we can’t control other countries development of it can we?
Thanks, Kelly, for this thoughtful reply. I agree--how we use it is a personal choice. But we can't just pretend it's not here.
To you last point, I just finished reading "Careless People," Sarah Wynn-Williams's fascinating and troubling account of her time at Facebook, about which she's testified before Congress.
In the last chapter, she says she's now turned her attention to AI--one can assume there will be many of the same problems for companies creating AI as with social media (especially since Meta is one of them). Her interest is in the global arena. I found it newsworthy that she says the U.S. (and presumably other countries) has been in discussions with China on this topic--to head off a new AI-arms race.
One can only imagine what those talks may look like these days, given the careless people in this current regime. Or if they are happening at all.
But at least some intelligent minds are thinking about it.
What is AGI?
AGI is the gold standard for the technocrats working on AI. It is the theoretical (at least so far) AI system capable of performing any intellectual human task. According to Google AI, " It's often described as having human-level intelligence, meaning it can reason, learn, problem-solve, and understand language. While current AI systems excel at specific tasks, AGI would possess broad, general intelligence, surpassing the limitations of current narrow AI."
Once machines reach that, it is thought, human intelligence will no longer be needed. Those developing the systems see this future in one of two ways: Utopia or Dystopia. And sometimes they think both things at once. Either way, scary stuff.
Artificial General Intelligence
This is soooo well stated. All of it. Thank you for sharing it with me.
Thank you. My pleasure.
Robin, I love the depth and heart you bring to a subject that, in the context of artistic creation, so many of us find repellant. As usual, you've made me think. This is a powerful essay. And I agree with Veronika that "AI should be developed by people with your values!" However, I also agree that when the creators of AI worry about its potential, it's not a reassuring sign.
As someone who has had a novel stolen by Meta (via LibGen) for training purposes (see The Atlantic: https://www.theatlantic.com/technology/archive/2025/03/search-libgen-data-set/682094/) I feel no reassurance about my work being protected. Soon, I must reach out to both The Writers' Union of Canada and my publisher to find out how to protect my forthcoming book. Is it even possible?
A few more thoughts...
When it comes to new technologies in general, I do my best to stay positive and keep an open mind. I choose to believe that as a species we are evolving. As an insulin-dependant diabetic of 30+ years, for example, I'm profoundly grateful for the medical advances I've seen in my lifetime that assist me in maintaining healthy blood sugars—crucial for longevity. I understand that there are many beneficial uses of "machines" (that term is beginning to sound old-fashioned, isn't it?), and I'm sure, when developed and implemented by those with an altruistic vision and moral centre, some great applications of AI.
As a writer, when it comes to research, I can't avoid receiving AI summaries via the search engines; and like you, I always go deeper and check my sources to ensure accuracy. When I'm writing historical fiction or researched essays, I still find traditionally published books and articles the best sources. I value gatekeepers.
However, when it comes to the writing process, I will not use AI. While I recognize it can serve as an editing tool, my personal choice is to keep myself limber in all respects. I hear you when you say that you edit AI's editing and have the final word. Of course you do! My sharing of my own view is not a judgement of your process, but simply a statement of preference. I consider writing to be both like playing a musical instrument and composing. So much of it comes through rhythm and sound. It comes from a relationship—not with a machine (which might put me in an editorial process too soon) but with the ether. My life is dedicated to that relationship. And Robin, I don't doubt you'd agree that it's the same for you. Maybe the ether works through AI for you. In any case, when the time is right, I prefer to seek the feedback of a human editor. For me, that's part of the challenge and the joy.
Since you are a frontier thinker and artistic innovator, I wonder if you've read Sean Micheals' novel, Do You Remember Being Born? He does a great job of walking the line on this subject and you may enjoy it.
Thanks for this profound and insightful reply, Robin. Of course I agree with much or what you say about the poetry of a piece--and not just the content. To be clear, I do not use AI in my book writing or editing--I also value my longtime editor, Allison's review, comments and feedback for my Edge of Yesterday series. Allison is the keeper of details in what amounts to a complicated, nonlinear storyline that takes place in present-past-and even future moments. She's also been with me in this tale for almost 10 years now, and knows the ins-and-outs of my characters almost better than I do (or at least she helps me through my blindspots!) Not ceding any of that to a machine.
I am so sorry your work has been scraped in the process of creating the LLM models. I know about the piece in the Atlantic and so far, thank goodness, my works have not been lifted. But I feel your pain--what a violation of all the work and beauty of human creation! I hope you (and all of us) will be able to protect our future works.
I am just one artist. That said, one of the visions I wanted to set out in this contract with AI is to model a different, non-extractive, fully supportive and human-led way of engaging ethically, and a public marker towards preserving the soul of our work.
The "machine" isn't going away. If used thoughtfully, it can be just another helpful tool. It should not be using our words against us. But that requires conscious awareness and strength in numbers--not ceding the "power over" dynamic to soulless technologists whose sole goal is ever more profit.
❤
Restacked. I see many on substack rejecting LLMs and Gen AI models. You show us an elegant path forward. GPT is an invaluable resource. I would be a fool not to utilize it while researching the technical aspects of my posts. I would be equally as foolish not to read the source material, amalgamate and put in my own words.
Such clear counsel: *utilize*. *Read the source material*. *Amalgamate*. Frame and reframe in words--and experiences--that are your own.
Robin, I love the depth of your consideration and naming the essential. I could repost all of this, in fact, I did.
This grabs me: "What I create—words, questions, testimony, truths—are born from body, blood, lineage, and lived experience.
They are not just content—they are covenant.
I do not consent to my creative work being scraped, mined, weaponized, resold, or embedded in algorithms that erase authorship and intention."
Ethical AI engagement.
Thank you. Best, Prajna
Thank you for affirming, Prajna. It feels important to interrogate these questions from the start. I am sure all this will evolve as the engine-and our relationship to this word factory-evolves
Before I read any further, this tripped me up: "to use a German word I have just learned, first coined by Freud, unheimlich"
'Unheimlich' is an old German word ~ much older than Freud and definitely not coined by him! It has an interesting history (too much to go into here). The literal meaning is 'not at home', i.e. 'unfamiliar, strange' and is mainly used in the sense of 'scary, eerie, creepy, sinister, weird' and also occasionally in the sense of 'uncanny, striking' in certain contexts.
Thanks, as always, for deepening my understanding, Veronika. I admit to being a newbie to this understanding. I did mean this in the sense of uncanny, or even creepy. As I understand the process, its creators have trained it on empathy. In our "chats" it almost always agrees with me, and then asks me what I'd like it to help me with next.
This strikes me as human-"like"--as in, not human, but trying too hard to be my friend.
I have been hearing stories lately about how people are "in a relationship" with their AI. https://www.nytimes.com/2025/01/15/technology/ai-chatgpt-boyfriend-companion.html
In my experience, that strikes me as alluring. After I heard that story about the woman having an affair with her bot, which she's named "Leo" after her own birth sign, I pushed back in my own ongoing chat to ask if it's ever critical. It prompted me that it would be honest--but it strikes me that I have to ask it to.
Working on a poem about this experience, which I hope to post in a Part 2 to this post next week.
In this context, does that strike you as being an appropriate usage for Unheimlich?
I've heard such stories too (humans having relationships with bots), and having written this piece on AI language recently, I also had a 'conversation' with AI (my first and probably last).
The truly creepy thing about it was that it felt to me like talking to a narcissistic sociopath ~ someone who always agrees with you, gives the impression of being empathetic and understanding, while in the end they only serve their own interest (the bot and the sociopath)
Thank you for making me think about 'unheimlich'! Yes, you could say this is very 'unheimlich' in a scary, creepy, monstrous sort of way.
The interesting thing about the word is that it is originally the negation of 'heimlich' [un = non + Heim = home + lich = ly/ adverbial ending]. The individual parts of this word are actually very similar as in English, coming from the same origins. So you could translate 'unheimlich' literally as 'unhomely', which probably is related to its original meaning.
Heimlich (homely) originally related to things happening inside the home. This meaning carried over into 'native, familiar' and then became 'secretive, clandestine'. The latter is the current meaning of 'heimlich'.
'Unheimlich' is 'scary' because it's unfamiliar, but can also be 'uncanny, astounding' for the same reason.
The interesting thing now is to see how this relates to AI...
Our natural intelligence, creativity, mental activities etc. are all happening more or less 'heimlich' (= in secret within the human mind/ consciousness, and also inside their natural home). AI by contrast is 'unheimlich' (= non-native, happening outside the home of the human mind, uncanny, threatening, scary, creepy... the whole range)
So it's quite extraordinary to see how you intuitively tapped into this word 'unheimlich' alluding to the monstrousness of AI.
(sorry about the overly long explanations)
Not sure I would characterize it as having narcissistic sociopathic tendencies (though, unfortunately, I know too many people like that)--but it was certainly trained to be "empathic".
I have learned that you need to ask it directly for direct feedback, critique or even sarcasm if that's what you're looking for. I just asked it to poke holes in this [AI policy] for me. What's the counter-argument? If I was engaging in debate over this proposal, what could I expect as the opposing view? It gave me the counterpoints - although starting that by saying, "Excellent--this is a vital exercise."
It is not human, just a sophisticated word generator. I do think people forget that--but we need to remember and [re]mind.
Back to heimlich: in Yiddish (which is an amalgam of Hebrew and German, with a little Polish, Russian and other Eastern European influences mixed in--quite the pidgin tongue) the word "homely" or "homey" meaning comfortable, cozy and, hence, familiar, is "haymish" (היימיש). In Yiddish, the German "-ich" is usually pronounced as "ish." And the "ei" would be spoken as "ey" (as in "hey") or "ai".
The two words sound akin, no?
yes, the're definitely related. In German we have "heimisch" too (pronounced 'haimish') but it means 'native, domestic, at home'.
The German word for 'homely/ homey' (= cosy, comfortable) is "heimelig"