You’ve packed a lot of meaning in a very short set of paragraphs. I’m glad I came back to reread it. (Read it twice this morning) 🙏🏽.
I’m sure you have mentioned elsewhere, but I think a lot of what is missing from the conversation around “computer-assisted writing” or “software assisted writing”, is that so much of the last 15 years has seen an increasing algorithm-zation of the presentation of all types of news, and obviously even the bodies of government.
But we should also at least step back, as you allude to in this brief post, and recognize that propaganda has not really ever *not* been with us …..
…..writing has seemed to become especially terrible in the last 15 years across many different formats as people seem to be pursuing the gamification of a New York Times bestseller and click bait headlines.
If software assisted writing is making these algorithms more efficient and maybe also slightly higher quality, we can only really stand on the sidelines and observe a lightening different icing on the pre-existing dystopic trend of “what is news” .
At the same time, if it pushes writers in general to be better writers, then that is something to be grateful for even if reluctantly.
Likewise, much easier said than done, but this is an absolutely phenomenal and extremely challenging opportunity for all high school and college and graduate school level professors to raise the standards of their engagement given the tools at our (inundated) disposal.
People/students “should” be able to offer and defend an absolute labyrinth/encyclopedi, set of sources and quality of prose.
On the opposite end, having people defend what they write from an oral presentation , without notes , is also a powerful opportunity, …..
but admittedly challenging to grade and time consuming.
(and what do you know, even I was able to write this without relying on a “software enhancement or editor,”…. Giving credit and acknowledgment to my partner for making muffins while I was reading and writing this, and my 2 1/2 year-old for jumping on my back, pretending to be a unicorn and feeding me cardboard cylinder, pretending to be a taco).
I'm glad this post landed with you. Your thoughts and reflections track a great many of my own, including the ones on education. I'm actually slated to lead my college's semi-annual all-faculty meeting two weeks from now, at which I've chosen AI as our central topic. My theme will be, without quite putting it this way, "Just what the hell are you all encountering in your classrooms and in your relationships with students when it comes to AI? How is this interacting with anything else? Where are you discouraged, encouraged, and totally at sea? Do you have any exciting things to report, or any frustrations and laments? What's working, what's not working, and what seems broken at this point?" This will be preceded by a presentation from me in which I sketch the state of AI in American higher ed right now with the help of info and perspectives from several carefully chosen articles and essays plus the crowd-sourced input from my faculty that I have already been gathering from them. I may say something here about what comes of all that.
Separately, your description of your child's antics as you typed your comment is delightful!
I think AI is changing faster than we have the capacity to adapt to it. It would take us five to ten years to fully absorb what AI can do right now, yet in five years, three years, or perhaps even one, AI is going to be radically more advanced than it is now. I think the answer for writers is, as you say, to hold steady to the fundamentals of human writing.
The likely pace of change -- hell, the *current* one -- is indeed dazzling, isn't it? Interesting how so many of us are presently recognizing that this all offers us the opportunity, if we want to recognize and use it as such, to clarify and maximize the sui generis human aspect that has always stood in back of our writing. And also the uniqueness each of us brings as an unrepeatable individual.
"But some of us are acting as if a technology that has only just been sprung upon us, one so young that we don’t even know what it is yet or what its lasting cultural meaning will be, has already changed the whole game."
For 3000 years, a piece of writing was evidence of human thought, and the fluency of that writing was a loose proxy for the quality (or at least degree) of thought that went into it. For the most part, good writing meant there was a mind behind the scenes that had done a lot of work.
How do you parse the relationship between thought and writing itself? I’m thinking of Plato’s myth/story in the Phaedrus of King Thamus and his rejection of the gift of writing that the god Theuth tried to give humankind. Thamus protested that externalizing thought like that would result in a loss and degradation of actual thought and memory as people outsourced such things to a tool, leading to a puffed up sense of false wisdom and knowledge. I’ve long found this myth to be usefully centering for my thinking about such things, a tool in itself for reminding me that writing itself, which has come to be the standard-bearer of what we widely consider to be education and educated knowledge, is actually an external tool, a technology, and that all the various specific techs that we’ve developed throughout history for practice it—from the stela, the scroll, and the codex to the printing press, the internet, and the tablet computer—are subsequent to this primary technology of writing itself. At the source of all these things lies the invention of the technology of writing itself, which we widely tend to forget is a technology, a tool with which we engage to enhance and even enable thought.
Prior to that invention, proof of thought resided in oral communication, and in bodily movements and gestures, and in manner and mode of clothing, and in cultural/societal agreements and practices, architectural arrangements, patterned gatherings and placements of sticks and stones, and so on.
I find myself reflecting back to the earliest beginnings of all these things—aided, of course and no doubt, by the fact that I’m literate in the technology of writing—to mull over the ways in which generative AI may interface with all of that.
You're right that the advent of the written word could arguably be classified as one of the first instances of cognitive "offloading." But I would like to look more closely at what is being offloaded. If you're looking at the end result (knowledge) then that is of course the case.
Look at the process itself, though, and writing is closer to verbal communication than we assume... both require a compression of ephemeral thought into linear narrative. AI has moved past the stored knowledge aspect and encroached on this compression, which is where the magic happens (in my humble opinion). And where it's lost, if you offload it.
A similar ploy is happening in music, with bands like Deerhoof writing music that is ‘inconsistent in style, hard to copy, every song different, and full of exactly what AI can't do: improvisation, surges in feeling, a sense of humor, and human (or animal) interaction’ (https://deerhoof.website/random-thoughts-on-ai-and-music/). This is primarily in response to the fact that AI systems such as Suno’s have been trained on vast swathes of essentially stolen data, and regurgitate almost sound for sound much of the music they’ve been fed. This is already having an enormous impact on the music industry, with upwards of 40% of tracks uploaded to streaming services every day being produced by AI. There would already seem to be a moral imperative to look at modes of resistance when seen from this perspective. I’ve also written elsewhere about my take on the spiritual damage AI-generated music can do (https://johngarner.co.uk/writings/the-spiritual-impacts-of-ai-on-music).
Although there are differences between the roll out of gen-AI in music and writing, presumably they aren’t miles apart? Has ChatGPT, for example, not been trained on data which has been scraped without consent (and likely loaded in by hidden labourers working in appalling conditions under authoritarian regimes)?
There is an urgency to this, in my opinion, which goes beyond the questions of the relationship that might ultimately unfold between humans and gen-AI in creativity.
As usual, brilliant observations and plenty of food for thought. I'll be reading it again!
The only antidote to machine-made art is more human-made art and I think we'd be better served as a species to just get on with that instead of all the hand-wringing and teeth-gnashing. Though that could be me opting for the ostrich option.
As for changing how we write to seem less AI-ish, well, you'll have to pry the em dash out of my cold, dead hands. :)
You’ve packed a lot of meaning in a very short set of paragraphs. I’m glad I came back to reread it. (Read it twice this morning) 🙏🏽.
I’m sure you have mentioned elsewhere, but I think a lot of what is missing from the conversation around “computer-assisted writing” or “software assisted writing”, is that so much of the last 15 years has seen an increasing algorithm-zation of the presentation of all types of news, and obviously even the bodies of government.
But we should also at least step back, as you allude to in this brief post, and recognize that propaganda has not really ever *not* been with us …..
…..writing has seemed to become especially terrible in the last 15 years across many different formats as people seem to be pursuing the gamification of a New York Times bestseller and click bait headlines.
If software assisted writing is making these algorithms more efficient and maybe also slightly higher quality, we can only really stand on the sidelines and observe a lightening different icing on the pre-existing dystopic trend of “what is news” .
At the same time, if it pushes writers in general to be better writers, then that is something to be grateful for even if reluctantly.
Likewise, much easier said than done, but this is an absolutely phenomenal and extremely challenging opportunity for all high school and college and graduate school level professors to raise the standards of their engagement given the tools at our (inundated) disposal.
People/students “should” be able to offer and defend an absolute labyrinth/encyclopedi, set of sources and quality of prose.
On the opposite end, having people defend what they write from an oral presentation , without notes , is also a powerful opportunity, …..
but admittedly challenging to grade and time consuming.
(and what do you know, even I was able to write this without relying on a “software enhancement or editor,”…. Giving credit and acknowledgment to my partner for making muffins while I was reading and writing this, and my 2 1/2 year-old for jumping on my back, pretending to be a unicorn and feeding me cardboard cylinder, pretending to be a taco).
I'm glad this post landed with you. Your thoughts and reflections track a great many of my own, including the ones on education. I'm actually slated to lead my college's semi-annual all-faculty meeting two weeks from now, at which I've chosen AI as our central topic. My theme will be, without quite putting it this way, "Just what the hell are you all encountering in your classrooms and in your relationships with students when it comes to AI? How is this interacting with anything else? Where are you discouraged, encouraged, and totally at sea? Do you have any exciting things to report, or any frustrations and laments? What's working, what's not working, and what seems broken at this point?" This will be preceded by a presentation from me in which I sketch the state of AI in American higher ed right now with the help of info and perspectives from several carefully chosen articles and essays plus the crowd-sourced input from my faculty that I have already been gathering from them. I may say something here about what comes of all that.
Separately, your description of your child's antics as you typed your comment is delightful!
Oof! I felt some sympathetic anxiety and envy hearing about that facilitated discussion. They’re lucky to have you leading it. 🙏🏽
I think AI is changing faster than we have the capacity to adapt to it. It would take us five to ten years to fully absorb what AI can do right now, yet in five years, three years, or perhaps even one, AI is going to be radically more advanced than it is now. I think the answer for writers is, as you say, to hold steady to the fundamentals of human writing.
The likely pace of change -- hell, the *current* one -- is indeed dazzling, isn't it? Interesting how so many of us are presently recognizing that this all offers us the opportunity, if we want to recognize and use it as such, to clarify and maximize the sui generis human aspect that has always stood in back of our writing. And also the uniqueness each of us brings as an unrepeatable individual.
"But some of us are acting as if a technology that has only just been sprung upon us, one so young that we don’t even know what it is yet or what its lasting cultural meaning will be, has already changed the whole game."
For 3000 years, a piece of writing was evidence of human thought, and the fluency of that writing was a loose proxy for the quality (or at least degree) of thought that went into it. For the most part, good writing meant there was a mind behind the scenes that had done a lot of work.
AI broke writing as proof-of-thought.
That’s a strong perspective/insight. Thank you.
How do you parse the relationship between thought and writing itself? I’m thinking of Plato’s myth/story in the Phaedrus of King Thamus and his rejection of the gift of writing that the god Theuth tried to give humankind. Thamus protested that externalizing thought like that would result in a loss and degradation of actual thought and memory as people outsourced such things to a tool, leading to a puffed up sense of false wisdom and knowledge. I’ve long found this myth to be usefully centering for my thinking about such things, a tool in itself for reminding me that writing itself, which has come to be the standard-bearer of what we widely consider to be education and educated knowledge, is actually an external tool, a technology, and that all the various specific techs that we’ve developed throughout history for practice it—from the stela, the scroll, and the codex to the printing press, the internet, and the tablet computer—are subsequent to this primary technology of writing itself. At the source of all these things lies the invention of the technology of writing itself, which we widely tend to forget is a technology, a tool with which we engage to enhance and even enable thought.
Prior to that invention, proof of thought resided in oral communication, and in bodily movements and gestures, and in manner and mode of clothing, and in cultural/societal agreements and practices, architectural arrangements, patterned gatherings and placements of sticks and stones, and so on.
I find myself reflecting back to the earliest beginnings of all these things—aided, of course and no doubt, by the fact that I’m literate in the technology of writing—to mull over the ways in which generative AI may interface with all of that.
Great reply!
You're right that the advent of the written word could arguably be classified as one of the first instances of cognitive "offloading." But I would like to look more closely at what is being offloaded. If you're looking at the end result (knowledge) then that is of course the case.
Look at the process itself, though, and writing is closer to verbal communication than we assume... both require a compression of ephemeral thought into linear narrative. AI has moved past the stored knowledge aspect and encroached on this compression, which is where the magic happens (in my humble opinion). And where it's lost, if you offload it.
A similar ploy is happening in music, with bands like Deerhoof writing music that is ‘inconsistent in style, hard to copy, every song different, and full of exactly what AI can't do: improvisation, surges in feeling, a sense of humor, and human (or animal) interaction’ (https://deerhoof.website/random-thoughts-on-ai-and-music/). This is primarily in response to the fact that AI systems such as Suno’s have been trained on vast swathes of essentially stolen data, and regurgitate almost sound for sound much of the music they’ve been fed. This is already having an enormous impact on the music industry, with upwards of 40% of tracks uploaded to streaming services every day being produced by AI. There would already seem to be a moral imperative to look at modes of resistance when seen from this perspective. I’ve also written elsewhere about my take on the spiritual damage AI-generated music can do (https://johngarner.co.uk/writings/the-spiritual-impacts-of-ai-on-music).
Although there are differences between the roll out of gen-AI in music and writing, presumably they aren’t miles apart? Has ChatGPT, for example, not been trained on data which has been scraped without consent (and likely loaded in by hidden labourers working in appalling conditions under authoritarian regimes)?
There is an urgency to this, in my opinion, which goes beyond the questions of the relationship that might ultimately unfold between humans and gen-AI in creativity.
As usual, brilliant observations and plenty of food for thought. I'll be reading it again!
The only antidote to machine-made art is more human-made art and I think we'd be better served as a species to just get on with that instead of all the hand-wringing and teeth-gnashing. Though that could be me opting for the ostrich option.
As for changing how we write to seem less AI-ish, well, you'll have to pry the em dash out of my cold, dead hands. :)