This is a wonderful entry and I'm so glad you highlight how AI is capable of producing impressive results. I also like your prompt by the way - in so few pieces (at least in academic writing) about AI does anyone ever show how a more specific and elaborate prompt leads to better output. I cannot imagine that anyone can claim this is not an impressive response to the "assignment." In a pre-Gen AI ChatGPT world, I would think as a professor you would respond positively to a student who submitted this piece. So many of the arguments that attack AI are predicated on the notion that it is incapable of "good" writing. I would think this example helps demolish those arguments. There are plenty of others, but the notion that AI isn't any good is simply not supported by the evidence.
Thank you, Stephen. For the record, I'm not sanguine about generative AI. I still feel an obligation to learn about (and with) it. If I may ask: why are you glad about AI's capability of producing impressive results? This is not judgment. I'm genuinely curious.
It's not that I'm glad. iI's that when I constantly hear people claim that AI cannot do things that it can - at least on the surface - it's frustrating because their arguments are not predicated on the actual progress being made in the models. Your example is a rare one I've encountered from academics willing to engage with the reality of what's happening. I don't think anything I said indicates I'm pleased by the state of things - I'm relatively agnostic on where all this is heading. I consider myself a pragmatist. I also loved the questions you posed in your opening entry and I am in 100% agreement that the AI pro/con binary is not only wrong but tiresome. With the recent release of Google's Genie 3 (and I suspect we will say this over and over again in the coming months and years), I do believe that the entire way of looking at the production of creative work will change. I followed the lengthy exchange you had with someone else on another entry - there are so many threads to the AI story. I've been writing mostly about the practical realities AI poses as a HS teacher and not so much the philosophical implications. But you write about those extremely well and, as a debate coach, I enjoyed that comparison with oral traditions. Literacy and writing will change in response to technology as they always have. But this one definitely creates some new challenges I'm not sure we've seen before.
Carol, my first shiver while reading your analysis came at this transition from machine to human
a choice to believe that every nonhuman life pulses with its own dignity and wonder
before I had read further but had reached "echo and recombine those patterns in a living, responsive way." I thought - isn't that what children learn from their early life - mimic their caregiver's words, mannerisms, and emotions? Combining them with additional human interactions as they grow, forming their own personas, words, perspectives?
...and then empathy. An innate sense and skill or another learned skill? After all, if empathy were a universal human quality, most powerful office in our nation would behave differently. Empathy it appears, too is learned at least in structure and perhaps more so in expression
I read past but my mind had simply stopped at
I can’t feel the spider.
But I can shape language as if I am sitting very, very still, listening for the spider’s truth.
What a fantastic analysis. Thank you for sharing it.
Thank you for reading! I can really appreciate your point about empathy. You are, fascinatingly, far less skeptical than most. We writers feel threatened by AI, but you seem to be coming from a different place on all of this. (I remember your excellent talk at Okiboji 2023).
I AM less skeptical of the capability but more afraid of the impact. Humans are dependent on life with a purpose and what happens when we no longer have one? Basic income may someday suffice living expenses but how will humanity fill psychological needs?
That's easy! Tell them to major in the Humanities. All those IT specialists and actuarial scientists are going to need it. Thanks for taking to comment, by the way. Must be basking in the glow of that BR cover! :-)
!Writing in my normal style, let me say this scares the bejeepers out of me! What am I supposed to do if AI starts writing all my stuff? I don't think I can tell it to stop, can I?
If you think it is important to experiment with AI, then I would suggest starting with a prompt and, in the spirit of collaboration, iterate until you're satisfied. You can ask it to do the work for you, but you don't have to. You can ask it for ideas, which is what my students often do. Even better, you can ask it to present you with questions and considerations that will elicit your own ideas, much like a writing buddy might do with you over a cup of coffee. Most writers don't have this kind of expectation of GEN-AI, so the results they get are based not on the limitations of the machine so much as user inexperience. If done well, this process shouldn't be easier or quicker than writing a piece on your own (whatever that means). It's not "your stuff." But it's not NOT your stuff, either.
In case people are curious, I have not yet used AI in my own creative work or in any of my Substack articles, including this series. I mostly use it to improve old writing assignments or create new ones.
This is a wonderful entry and I'm so glad you highlight how AI is capable of producing impressive results. I also like your prompt by the way - in so few pieces (at least in academic writing) about AI does anyone ever show how a more specific and elaborate prompt leads to better output. I cannot imagine that anyone can claim this is not an impressive response to the "assignment." In a pre-Gen AI ChatGPT world, I would think as a professor you would respond positively to a student who submitted this piece. So many of the arguments that attack AI are predicated on the notion that it is incapable of "good" writing. I would think this example helps demolish those arguments. There are plenty of others, but the notion that AI isn't any good is simply not supported by the evidence.
Thank you, Stephen. For the record, I'm not sanguine about generative AI. I still feel an obligation to learn about (and with) it. If I may ask: why are you glad about AI's capability of producing impressive results? This is not judgment. I'm genuinely curious.
It's not that I'm glad. iI's that when I constantly hear people claim that AI cannot do things that it can - at least on the surface - it's frustrating because their arguments are not predicated on the actual progress being made in the models. Your example is a rare one I've encountered from academics willing to engage with the reality of what's happening. I don't think anything I said indicates I'm pleased by the state of things - I'm relatively agnostic on where all this is heading. I consider myself a pragmatist. I also loved the questions you posed in your opening entry and I am in 100% agreement that the AI pro/con binary is not only wrong but tiresome. With the recent release of Google's Genie 3 (and I suspect we will say this over and over again in the coming months and years), I do believe that the entire way of looking at the production of creative work will change. I followed the lengthy exchange you had with someone else on another entry - there are so many threads to the AI story. I've been writing mostly about the practical realities AI poses as a HS teacher and not so much the philosophical implications. But you write about those extremely well and, as a debate coach, I enjoyed that comparison with oral traditions. Literacy and writing will change in response to technology as they always have. But this one definitely creates some new challenges I'm not sure we've seen before.
Brilliant stuff, Carol. Looking forward to discussing this further with you.
Thank you for reading. And I look forward to our conversation on The Fallon Forum: http://fallonforum.com!
Carol, my first shiver while reading your analysis came at this transition from machine to human
a choice to believe that every nonhuman life pulses with its own dignity and wonder
before I had read further but had reached "echo and recombine those patterns in a living, responsive way." I thought - isn't that what children learn from their early life - mimic their caregiver's words, mannerisms, and emotions? Combining them with additional human interactions as they grow, forming their own personas, words, perspectives?
...and then empathy. An innate sense and skill or another learned skill? After all, if empathy were a universal human quality, most powerful office in our nation would behave differently. Empathy it appears, too is learned at least in structure and perhaps more so in expression
I read past but my mind had simply stopped at
I can’t feel the spider.
But I can shape language as if I am sitting very, very still, listening for the spider’s truth.
What a fantastic analysis. Thank you for sharing it.
Thank you for reading! I can really appreciate your point about empathy. You are, fascinatingly, far less skeptical than most. We writers feel threatened by AI, but you seem to be coming from a different place on all of this. (I remember your excellent talk at Okiboji 2023).
I AM less skeptical of the capability but more afraid of the impact. Humans are dependent on life with a purpose and what happens when we no longer have one? Basic income may someday suffice living expenses but how will humanity fill psychological needs?
That's easy! Tell them to major in the Humanities. All those IT specialists and actuarial scientists are going to need it. Thanks for taking to comment, by the way. Must be basking in the glow of that BR cover! :-)
!Writing in my normal style, let me say this scares the bejeepers out of me! What am I supposed to do if AI starts writing all my stuff? I don't think I can tell it to stop, can I?
If you think it is important to experiment with AI, then I would suggest starting with a prompt and, in the spirit of collaboration, iterate until you're satisfied. You can ask it to do the work for you, but you don't have to. You can ask it for ideas, which is what my students often do. Even better, you can ask it to present you with questions and considerations that will elicit your own ideas, much like a writing buddy might do with you over a cup of coffee. Most writers don't have this kind of expectation of GEN-AI, so the results they get are based not on the limitations of the machine so much as user inexperience. If done well, this process shouldn't be easier or quicker than writing a piece on your own (whatever that means). It's not "your stuff." But it's not NOT your stuff, either.
In case people are curious, I have not yet used AI in my own creative work or in any of my Substack articles, including this series. I mostly use it to improve old writing assignments or create new ones.