Like Sci-Fi writer Connie Willis, I approach new tech with, "What could possibly go wrong?"
Frankly, what worries me most about AI is that its developers are in such a tremendous hurry. They all want to be the indispensable whatever, without much clarity about what they are creating and foisting on humanity. This mad rush marginalizes questions - about unintended effects, about ethical issues, about utility vs "we have it so we must use it, in everything we do, all the time" - that surely we should be asking about a world- rattling tool.
My Luddite heart is deeply concerned about waste of resources (water especially in the arid - flammable - west, but also the surge of energy required to run these data centers) and the difficulty of using internet tools without it. I tried turning off the AI enhancements on my phone with zero success - the search function in settings does not acknowledge "turn off AI".
I don't use it. I'm at an age when I derive pleasure from using language skills acquired over many decades of reading & writing & wordplay, saying what I want, the way I want to.
Fair enough. I feel like you do in regard to my many decades of reading and writing and wordplay and creativity. And yet I use AI almost every day, just not for those purposes. There will be drawbacks and there will be progress. I tend to look at new technologies through the examples exhibited by old ones. Take the automobile. When it first came into widespread use not too many people thought about the problems it might cause. It meant improved movement and freedom and greater productivity. But it also gave us traffic nightmares and pollution. AI will mean the same combination, we just don't know the exact details yet.
The binary nature of the AI debate is infused with a sense of existential loss. The Luddites showed us how that went 200 years ago. It makes perfect sense because AI - we do not have AI and never will through LLMs, though they are used interchangeably - promoters have told us it will completely reconfigure industries which means capital wins in this version of history and labor, writers and artists lose.
This message is echoed to us through the mainstream business media who have taken the side of big tech and capital. I have an idea for a book that will investigate how it is that established media decided to be uncritical of anything to do with AI but that can wait till after the crash.
One thing to observe about LLM use to date is that it has been subsidized and as corporations switch off using it because it's expensive and delivers little so will others when market pricing makes the personal indifference curve decline using it. That won;t mean it goes away; it may be better in future, it may be just another convenience tool like excel macros or whatever and as we know how technology transforms creation, production and exchange modification is inevitable, though not as its been pitched by the vocal advocates thus far.
I tend to think that the advancement of AI will be closely tied to its ability to generate profit for those who use it. For example, you mentioned that companies and corporations are switching off from it because of the cost and the small return. I bet that will mean they encourage their employees to use it at their own expense. They will put the burden on their employees, telling them that they need to increase production but that the company cannot afford it. And employees will do that because they will think it gives them a chance to get ahead in the competition for advancement.
That may occur depending on the territory and how labor laws deal with it. Contractors in IT would face that issue. Anyway, just to let you know, I made an effort to be on Substack and I’ll be doing more on AI and topics from my consulting background as they relate to writing.
Joe, Friends: For all who wish to broaden their knowledge and understanding regarding the much debated AI issues, I urge you to read Javier Ergqueta’s latest posting in Light, Not Heat, entitled “Should we step on the AI accelerator? Or the brake?”
Put the following in your browser. I promise you won't be disappointed
Javier Ergqueta is an international educator, researcher and developer of critical thinking methodologies (and a personal friend whom I’ve long admired for his brilliance, his reasoning, and his creative gifts.)
Well said: "LLMs are tools that rely primarily on compilations of standard, mainstream literature, and I don’t want to write like that. They mimic writing styles, rather than attempt to create anything imaginative or original." A nicely balanced post overall.
I won't claim to always perfectly follow my own advice, but I try my best.
I used a little bit of generative AI for images when Midjourney first started the trend. I was mostly curious. But then I saw a very angry post on BlueSky (where else?) from an illustrator that convinced me that it's no better for me to do that than to use AI to generate my own writing. And I've always been incredibly appreciative of the cover artist for my novel, who obviously went through the trouble of reading a publisher's galley of it before she sent off comps to me to approve.
Also, I like playing around with Photoshop. I've been using it since it was a tool called Barneyscan bundled with scanning software. There's really no good excuse for me using AI in my Substack art.
But here's the catch. Sometimes I do, anyway, because I use Adobe Stock as a base for a lot of my images. But they now include AI art in their stock images. They claim that the stock AI is all produced by LLMs that compensate illustrators that provide the stock images to Adobe or through its formerly named Behance portfolio library (not sure what that's called these days).
So if I forget to add a filter to my stock search or don't look for the tiny "AI" icon in the search results, it's possible that my caption, "base image licensed from Adobe Stock" is tainted. I should really look for another stock service that is free of AI. I haven't researched that, though.
As for AI writing, hell, I've been using Grammarly since before AI was a term, almost, but I've never trusted it much. I really just use it as a spell checker, and even then it screws up. As for "writing" with AI, I have so many ideas bouncing in my head that the last thing I need is AI entering my neuron zone and interfering with my writing process.
My opinion about that is pretty strong. If you need AI to help you generate writing ideas, why are you calling yourself a writer? Trust me, there are MUCH better ways to make a dollar.
The impression I get from what you have said here is to point to the increasing ethical dilemma of using artificial intelligence as it continues to infiltrate every aspect of the creative life. See my response to Deb Levy, above. I will be working on a post about that experience that I hope to publish in the next couple of weeks. And I'd be very interested to get your opinion about what I'm doing.
I'll look forward to that. I'm a software engineer as well as a writer, but mostly retired now, so my primary experience in AI-related software coding is through tinkering with Claude. It will be interesting to see what someone who isn't a software engineer does with it. I've read some about that, but this will feel closer to home.
The reaction I'm seeing from my former (mostly) colleagues in the biz is mixed. James Gosling, the inventor of the Java programming language, writes on LinkedIn that AI gives engineers the headaches of software engineering (bugs, maintenance nightmares, spaghetti code), while taking away the part we love the most, which is the creative effort of producing stuff with code. In other words, we now get to deal with the cruft and headaches, while the AI tools get to have all the fun. Others, however, say that properly used, AI can save hundreds of hours of time and allow for more creativity, not less.
Most engineers I know of speak from the perspective of a large code base, as opposed to smaller apps that folks like you might be more interested in.
Good morning Joe. To me, there are two extremes: people who enjoy working things out for themselves and those who can’t/won’t and want/need technology to do it for them. AI is a useful tool…if used properly; dangerous, if not. In creativity, I think it will take most, if not all, the pleasure one derives from doing it yourself, the result being you might as well fold your tent and be a couch potato. Personally, I will never use it as I derive too much pleasure in thinking about what I plan to write and sitting in my chair with my chickens and ducks around me, figuring out how to write an entertaining story.
Thanks, Joe, for broaching this difficult subject; it's a ball that's going to be batted around for some time to come. There's a lot I want to remark on, but I'll confine myself to a few items
I note (and I'll note it for myself, as well) that it seems necessary that we must make our mark in some following way: I want to talk about AI but let me assure you I don't use it for writing. It's like a badge of honor we need to pin to ourselves before the discussion can continue.
But how can we know who uses and doesn't, and to what extent? Humans don't like cheaters (likely an inborn trait). We'd have no qualms about calling out someone who used an AI-enhanced baseball bat (or used performance-enhancing drugs) and proceeded to knock the ball out of the park every time because we believe one's talent should be the mark of excellence.
How much fudging do we allow before AI-enhanced writing is called out? Or, because it's hidden somewhat, is there any point to calling it out? One commenter asked a really good question: when the AI gives "advice" and you take that advice (because it's good advice?) at what point have you given up your talent to the machine? Moreover, if you didn't know the advice beforehand, then how would you know it's good advice, anyway? You'd have to trust the AI--another level completely--as you would trust a human developmental editor. But on what basis?
I'm practical, all the same. History shows that if an advantage is offered, people will take it, and those who don't will fall far behind and away, banished to quaint artisan woodworking shows people watch on TV. AI is here to stay, in some form or other, and incrementally it's going to play an ever-larger role in all creative endeavors. Ultimately, each will have to decide, based on their integrity, how they go about using AI.
As God surely said to Adam, "This is Earth. Purists need not apply."
See my response to Deb Levy, above. Based on what you've said here, should I be trying to code an app using AI? Have I crossed some ethical line in this pursuit? And if not, then why shouldn't non-writers, or bad writers, use AI? This is the question I am going to try to explore in a couple more weeks.
I'm really looking forward to more of your thoughts on AI and those of commenters; Charles Bastille made me go look up a lot about data centers.
If you are a non-coder and you ask AI to code something based on your description, you should be proud of the description if it does it well. You aren't up for a National Coder Award, so your coding skills aren't judged. No issue.
If you are a coder and you can't figure out how to code something like an improvement on an app, you probably shouldn't submit your name for said award. And I suspect pride's opposite would be the feeling.
There are a lot of related issues. For example, AI as one's developmental and line editor reminds me of the Carver-Lish controversy; who deserves credit? I don't know how much influence AI can have before a writer will admit that it was mostly AI, but I suspect there'll be no admission, unless caught, then ...
I played with AI for a while (on old stories I thought were lousy), and I was very impressed, but I stopped when I realized that the painfully fun feeling of struggling to write a damn good sentence was gone, when it had been made so easy by AI.
I can go out to a restaurant for dinner, or I can prepare it myself from scratch. Guess which version is more satisfying? I think you're onto something when you mentioned the painfully fun feeling that goes with writing well. In the not too distant future everyone will be an adequate writer using AI. But those of us who continue to do it ourselves will earn a feeling that the rest cannot imagine.
For years, I was in a weekly critique group and, though I learned a lot by giving other writers free & extensive feedback, it was a heavy time commitment. I no longer have all this time to spare. There is a lit-mag site where they added an A.I. critique tool; it can be dialed all the way up (harsh, brutal feedback) or backed down to light or medium critique. I paste in my own text; I get a free critique back. Then I go on with my life. It is a labor-saving device like any other.
Joe, I think I’ve said this before (and maybe before that too), but I always look forward to reading your posts, which are thoughtful, well-balanced, and always make me think. And speaking of thinking, I’ve been doing a lot of it re: AI. You make an excellent point that the controversy surrounding the new technology, those for and those against (Creative Resistance) may be a social issue viewed through the lens of politics. Amen to that. Personally, I don’t know where I stand, but I’ve used AI, it’s amazing in its own right, and as far as flagrantly letting AI write a story, book etc., I think that’s totally wrong. However, how is having an editor/beta reader/friend give editorial suggestions/advice, and the writer taking that advice and changing/rewriting his/her work any different than taking the suggestions of AI? That’s the argument I’d like to have explained to me.
I can't answer that question, but I am working on something that I find very interesting and hope to write about in the coming weeks. I am using AI to try to write an app. It's nothing major, just an improvement on something that some friends and I use online. But I have been following the advancements in AI and noticed that you can now use it to describe how you want your app to work in plain language, and AI will code it for you. I cannot write code. I cannot even begin to imagine how to write code. But I have been working with AI for a couple of weeks now and we are making progress. And in a few more weeks, I hope to have a working application. I will get into more details about how fascinating this process has been. But what I want to say here is that AI is giving this non-coder an opportunity to do something he could never do before. I will never replace a real, experienced, imaginative coder. And that's the similarity I see to the writing world. AI will hopefully never replace experienced and imaginative writers. It will give people who really can't write very well the opportunity to do something they can be, frankly, proud of. Believe me, if I can get this app to work I will be as proud of the effort as I have ever been with any of my writing. What does all this mean? I am thinking through the implications.
Deb, AI can be a time-saver. EX: My historical essay on the New England Vampire Panic was approx 1,500 words. After I created a "vampire panic" video, I asked AI to condense my own essay to 300 words to use for my YouTube description. This task would have -- in the past -- taken me no less than an hour. I am happy to assign all my scut work to an uncomplaining AI assistant.
From an ethical standpoint, that's a good observation, but I personally wouldn't trust AI to make editorial recommendations unless I was just looking for obvious, glaring plot holes. I'd much prefer a human beta reader (who I also might ignore). The way I see it, I don't think Dickens or Dostoevsky or the lesser writers of their eras had beta readers. I've been reading literature since I was in third grade. I trust myself a lot more than I trust a lot of others, human or otherwise. This is likely a flaw, but it's my flaw.
I think using it for research is fine. I sometimes use it as a launching pad, then follow up with a heavy duty validation on my own. Sort of like checking the work of someone else's research. But it saves a lot of time, especially if there's a ton of research involved.
I love free A.I. help about what SEO words to include in my press release, summary, etc. 🥩 🥓 Additionally, I am doing far less harm to the world than people who eat meat or pollute the oceans.
Seems like there is something else the article merely alludes to. That is these data centers. I am one of the ones calling for regulation, not because I dislike AI, because I find it quite useful. Like you, I have no need of it in drafting or editing anything. I have decades of experience in creative and technical writing, and in generating compelling visual presentations. AI just isn't really smart yet, but it is fine as an aid to research and to generate some images. If you want labels in the images you really have to work with it so they aren't garbled.
But here is the regulatory issue: It is not just creative use that ought to be regulated, but data centers themselves. Perhaps it is a separate issue. If you think so, I get it. But consider the fights against the construction of new data centers. The problem in the US with that is that we have long been wedged apart by either/or logical fallacies. For example, big oil has told us for years we can either have good jobs OR clean water and air, but not both.
Ah, but why not both? A data center will consume upwards of five million gallons of water a day because massive amounts of water are required for cooling. Most data centers lose 70-80% of the water used in cooling. Some are constructed with better science in mind, and use closed loop cooling systems and recycle wastewater. Others, located in colder climes, can supplement water with air cooling.
What is happening that requires regulation is procedural. For decades in the US, this false choice between jobs and the environment has created a whole lot of ignorance. So, we have local governments who are okaying the construction of data centers without any public meetings, without having anyone else at the table to ensure the incoming centers will address water consumption AND electricity consumption. A proposed data center in Box Elder County, UT, for example, was going to over double the electricity consumption in the entire state! Seriously.
And, silly people, they were concerned about the depletion of their Salt Lake watershed over time and that their electrical bills would increase much faster than inflation. So there was a massive public outrcy and the developer had to back off. They didn't have the right people at the table, you see, and made the ill conceived choice to merely approve the new center administratively.
This also begs another question, Joe, and this is why we need all these new data centers. What are they really for? Why do we have this feverish need to build, build, build? What state are we hoping to achieve with this proliferation of data centers? Because they arguably are parasitical in terms of water and power usage UNLESS the approach to their location and building can be reasonably regulated to ensure closed loop cooling systems, recycling of wastewater, and leveraging solar/wind power to offset electricity consumption. There is even talk of mini fusion reactors to power centers, and those are expected to be grid-ready by 2030.
The bottom line is that progress is what it is. We cannot stop it, but we can ask why, demand purpose (if $X is invested in a new data center, our water supply will do Y, and the electricity consumption will go to Z, offset by A and B renewable energy sources). The problem with the United States is that we have become so very adversarial about everything.
But the question remains - Why do we need all these new data centers? How do they fit into overall sustainable economic development, urban planning and workforce development?
You make an excellent point. I did, in fact, gloss over the issue of data centers. It's kind of an out of sight, out of mind thing. There are no data centers where I live. Like so many other negative aspects of our society, they are relegated to areas in which the people have little power to fight them. But their effect on the environment is already profound, and probably will continue to get worse until more responsible aspects of science catch up. It would be great if the profit oriented segments of our society could work in tandem with the environmentally conscious ones, but we all know it doesn't work that way. The best we can hope for, since as you mentioned we are not going to stop AI progress, is that we can catch up quickly. A significant part of that is educating the population about the pros and cons of AI and the effect that progress is having on the environment. I've noticed that there really isn't much in major media about data centers.
RE Data centers: Read Will Lockett or Cory Doctorow (he's the enshittification guy) on the AI bubble. When it pops, it's going to leave a lot of empty data centers behind in its wake.
Pulitzer Prize winning novelist Colson Whitehead wrote, “People are usually surprised when I admit that I love AI. I use AI for everything, 24/7. I can’t get enough, me and my whole family.” He just doesn’t use it for his writing. He hastened to add, “Except art. If you use it for your art, you’re a freakin’ hack.”
I appreciate the points argued in this article. It prompts me to reflect on my negative response to last post’s AI image. I didn’t try to rationalise it at the time, but I did choose to share it. As a caveat, I’m not American.
As AI slop increases, I understand there to be a greater role for gatekeepers or curators. There are going to be decent things produced with the help of AI and a lot a lot a lot of subpar things. In the case of the image, perhaps it was that you had selected it and it felt so devoid of effort, the kind of clearly low quality AI aesthetic that makes up so much slop.
Interestingly, before AI could produce images even this well, I enjoyed the somewhat psychedelic images that a data scientist produced using, I think, GANs, some time in 2019ish. Again there was a level of curation though, as he had chosen only his favourites to share, printing and handing them out to us on cards.
You mentioned that it will be a greater role for gatekeepers and curators. Quite a few columnists (some mentioned in the post) have noted that the gap between people who know how to use AI and those who don't is just going to keep getting bigger. Let's hope that those gatekeepers and curators are among those who have the knowledge and experience to use AI for its best purposes.
A.I. helps save my sanity and conserve my energy for "paying markets" and their requirements. WHY? Because there is an entire eco-system of whimsical lit-mags who do NOT pay but who are still determined to suck the marrow from a writer's bones. EX: the lit-mag editor who accepts my CNF and then requests an essay on my inspiration. Or the lit-mag editor who accepts a poem and then demands an "author's note." Blah-blah-blah. Now I grind my WORD doc into A.I. + instructions & it spits out my 200-word "author's note." . . . . . . . . . . . . This gives me more time to devote to zines that actually pay poets and authors. ✒️ 🙏 ‼️ Shame on anyone who exploits another writer by denying payment.
Excellent point. I suspect the lit mag editors who do these kinds of things oh sincere, but it also shows how naïve, and frankly, self-centered they are. They think they are serving the "art," but they fail to see the bigger picture.
Lit-mags and zines that compensate poets and writers deserve my best. Those who do not pay and impose extra drudgery.... I give that to my "assistant" now. LOL
Like Sci-Fi writer Connie Willis, I approach new tech with, "What could possibly go wrong?"
Frankly, what worries me most about AI is that its developers are in such a tremendous hurry. They all want to be the indispensable whatever, without much clarity about what they are creating and foisting on humanity. This mad rush marginalizes questions - about unintended effects, about ethical issues, about utility vs "we have it so we must use it, in everything we do, all the time" - that surely we should be asking about a world- rattling tool.
My Luddite heart is deeply concerned about waste of resources (water especially in the arid - flammable - west, but also the surge of energy required to run these data centers) and the difficulty of using internet tools without it. I tried turning off the AI enhancements on my phone with zero success - the search function in settings does not acknowledge "turn off AI".
I don't use it. I'm at an age when I derive pleasure from using language skills acquired over many decades of reading & writing & wordplay, saying what I want, the way I want to.
Fair enough. I feel like you do in regard to my many decades of reading and writing and wordplay and creativity. And yet I use AI almost every day, just not for those purposes. There will be drawbacks and there will be progress. I tend to look at new technologies through the examples exhibited by old ones. Take the automobile. When it first came into widespread use not too many people thought about the problems it might cause. It meant improved movement and freedom and greater productivity. But it also gave us traffic nightmares and pollution. AI will mean the same combination, we just don't know the exact details yet.
The binary nature of the AI debate is infused with a sense of existential loss. The Luddites showed us how that went 200 years ago. It makes perfect sense because AI - we do not have AI and never will through LLMs, though they are used interchangeably - promoters have told us it will completely reconfigure industries which means capital wins in this version of history and labor, writers and artists lose.
This message is echoed to us through the mainstream business media who have taken the side of big tech and capital. I have an idea for a book that will investigate how it is that established media decided to be uncritical of anything to do with AI but that can wait till after the crash.
One thing to observe about LLM use to date is that it has been subsidized and as corporations switch off using it because it's expensive and delivers little so will others when market pricing makes the personal indifference curve decline using it. That won;t mean it goes away; it may be better in future, it may be just another convenience tool like excel macros or whatever and as we know how technology transforms creation, production and exchange modification is inevitable, though not as its been pitched by the vocal advocates thus far.
I tend to think that the advancement of AI will be closely tied to its ability to generate profit for those who use it. For example, you mentioned that companies and corporations are switching off from it because of the cost and the small return. I bet that will mean they encourage their employees to use it at their own expense. They will put the burden on their employees, telling them that they need to increase production but that the company cannot afford it. And employees will do that because they will think it gives them a chance to get ahead in the competition for advancement.
That may occur depending on the territory and how labor laws deal with it. Contractors in IT would face that issue. Anyway, just to let you know, I made an effort to be on Substack and I’ll be doing more on AI and topics from my consulting background as they relate to writing.
Joe, Friends: For all who wish to broaden their knowledge and understanding regarding the much debated AI issues, I urge you to read Javier Ergqueta’s latest posting in Light, Not Heat, entitled “Should we step on the AI accelerator? Or the brake?”
Put the following in your browser. I promise you won't be disappointed
https://substack.com/@lightnotheat
Javier Ergqueta is an international educator, researcher and developer of critical thinking methodologies (and a personal friend whom I’ve long admired for his brilliance, his reasoning, and his creative gifts.)
Well said: "LLMs are tools that rely primarily on compilations of standard, mainstream literature, and I don’t want to write like that. They mimic writing styles, rather than attempt to create anything imaginative or original." A nicely balanced post overall.
For those of you who are interested in AI-free images, I wrote a comprehensive guide on how to use images on Substack that respects copyright here:
https://www.ruminato.com/p/the-ruminato-guide-to-using-images
I won't claim to always perfectly follow my own advice, but I try my best.
I used a little bit of generative AI for images when Midjourney first started the trend. I was mostly curious. But then I saw a very angry post on BlueSky (where else?) from an illustrator that convinced me that it's no better for me to do that than to use AI to generate my own writing. And I've always been incredibly appreciative of the cover artist for my novel, who obviously went through the trouble of reading a publisher's galley of it before she sent off comps to me to approve.
Also, I like playing around with Photoshop. I've been using it since it was a tool called Barneyscan bundled with scanning software. There's really no good excuse for me using AI in my Substack art.
But here's the catch. Sometimes I do, anyway, because I use Adobe Stock as a base for a lot of my images. But they now include AI art in their stock images. They claim that the stock AI is all produced by LLMs that compensate illustrators that provide the stock images to Adobe or through its formerly named Behance portfolio library (not sure what that's called these days).
So if I forget to add a filter to my stock search or don't look for the tiny "AI" icon in the search results, it's possible that my caption, "base image licensed from Adobe Stock" is tainted. I should really look for another stock service that is free of AI. I haven't researched that, though.
As for AI writing, hell, I've been using Grammarly since before AI was a term, almost, but I've never trusted it much. I really just use it as a spell checker, and even then it screws up. As for "writing" with AI, I have so many ideas bouncing in my head that the last thing I need is AI entering my neuron zone and interfering with my writing process.
My opinion about that is pretty strong. If you need AI to help you generate writing ideas, why are you calling yourself a writer? Trust me, there are MUCH better ways to make a dollar.
The impression I get from what you have said here is to point to the increasing ethical dilemma of using artificial intelligence as it continues to infiltrate every aspect of the creative life. See my response to Deb Levy, above. I will be working on a post about that experience that I hope to publish in the next couple of weeks. And I'd be very interested to get your opinion about what I'm doing.
I'll look forward to that. I'm a software engineer as well as a writer, but mostly retired now, so my primary experience in AI-related software coding is through tinkering with Claude. It will be interesting to see what someone who isn't a software engineer does with it. I've read some about that, but this will feel closer to home.
The reaction I'm seeing from my former (mostly) colleagues in the biz is mixed. James Gosling, the inventor of the Java programming language, writes on LinkedIn that AI gives engineers the headaches of software engineering (bugs, maintenance nightmares, spaghetti code), while taking away the part we love the most, which is the creative effort of producing stuff with code. In other words, we now get to deal with the cruft and headaches, while the AI tools get to have all the fun. Others, however, say that properly used, AI can save hundreds of hours of time and allow for more creativity, not less.
Most engineers I know of speak from the perspective of a large code base, as opposed to smaller apps that folks like you might be more interested in.
Substitute the industry specific terms and this sounds a lot like what experienced writers say about AI use. Should be a very interesting exploration.
Good morning Joe. To me, there are two extremes: people who enjoy working things out for themselves and those who can’t/won’t and want/need technology to do it for them. AI is a useful tool…if used properly; dangerous, if not. In creativity, I think it will take most, if not all, the pleasure one derives from doing it yourself, the result being you might as well fold your tent and be a couch potato. Personally, I will never use it as I derive too much pleasure in thinking about what I plan to write and sitting in my chair with my chickens and ducks around me, figuring out how to write an entertaining story.
Thanks, Joe, for broaching this difficult subject; it's a ball that's going to be batted around for some time to come. There's a lot I want to remark on, but I'll confine myself to a few items
I note (and I'll note it for myself, as well) that it seems necessary that we must make our mark in some following way: I want to talk about AI but let me assure you I don't use it for writing. It's like a badge of honor we need to pin to ourselves before the discussion can continue.
But how can we know who uses and doesn't, and to what extent? Humans don't like cheaters (likely an inborn trait). We'd have no qualms about calling out someone who used an AI-enhanced baseball bat (or used performance-enhancing drugs) and proceeded to knock the ball out of the park every time because we believe one's talent should be the mark of excellence.
How much fudging do we allow before AI-enhanced writing is called out? Or, because it's hidden somewhat, is there any point to calling it out? One commenter asked a really good question: when the AI gives "advice" and you take that advice (because it's good advice?) at what point have you given up your talent to the machine? Moreover, if you didn't know the advice beforehand, then how would you know it's good advice, anyway? You'd have to trust the AI--another level completely--as you would trust a human developmental editor. But on what basis?
I'm practical, all the same. History shows that if an advantage is offered, people will take it, and those who don't will fall far behind and away, banished to quaint artisan woodworking shows people watch on TV. AI is here to stay, in some form or other, and incrementally it's going to play an ever-larger role in all creative endeavors. Ultimately, each will have to decide, based on their integrity, how they go about using AI.
As God surely said to Adam, "This is Earth. Purists need not apply."
See my response to Deb Levy, above. Based on what you've said here, should I be trying to code an app using AI? Have I crossed some ethical line in this pursuit? And if not, then why shouldn't non-writers, or bad writers, use AI? This is the question I am going to try to explore in a couple more weeks.
I'm really looking forward to more of your thoughts on AI and those of commenters; Charles Bastille made me go look up a lot about data centers.
If you are a non-coder and you ask AI to code something based on your description, you should be proud of the description if it does it well. You aren't up for a National Coder Award, so your coding skills aren't judged. No issue.
If you are a coder and you can't figure out how to code something like an improvement on an app, you probably shouldn't submit your name for said award. And I suspect pride's opposite would be the feeling.
There are a lot of related issues. For example, AI as one's developmental and line editor reminds me of the Carver-Lish controversy; who deserves credit? I don't know how much influence AI can have before a writer will admit that it was mostly AI, but I suspect there'll be no admission, unless caught, then ...
I played with AI for a while (on old stories I thought were lousy), and I was very impressed, but I stopped when I realized that the painfully fun feeling of struggling to write a damn good sentence was gone, when it had been made so easy by AI.
I can go out to a restaurant for dinner, or I can prepare it myself from scratch. Guess which version is more satisfying? I think you're onto something when you mentioned the painfully fun feeling that goes with writing well. In the not too distant future everyone will be an adequate writer using AI. But those of us who continue to do it ourselves will earn a feeling that the rest cannot imagine.
For years, I was in a weekly critique group and, though I learned a lot by giving other writers free & extensive feedback, it was a heavy time commitment. I no longer have all this time to spare. There is a lit-mag site where they added an A.I. critique tool; it can be dialed all the way up (harsh, brutal feedback) or backed down to light or medium critique. I paste in my own text; I get a free critique back. Then I go on with my life. It is a labor-saving device like any other.
Nyet for me. I pass. It’s either you know what you’re doing or you don’t.
Joe, I think I’ve said this before (and maybe before that too), but I always look forward to reading your posts, which are thoughtful, well-balanced, and always make me think. And speaking of thinking, I’ve been doing a lot of it re: AI. You make an excellent point that the controversy surrounding the new technology, those for and those against (Creative Resistance) may be a social issue viewed through the lens of politics. Amen to that. Personally, I don’t know where I stand, but I’ve used AI, it’s amazing in its own right, and as far as flagrantly letting AI write a story, book etc., I think that’s totally wrong. However, how is having an editor/beta reader/friend give editorial suggestions/advice, and the writer taking that advice and changing/rewriting his/her work any different than taking the suggestions of AI? That’s the argument I’d like to have explained to me.
I can't answer that question, but I am working on something that I find very interesting and hope to write about in the coming weeks. I am using AI to try to write an app. It's nothing major, just an improvement on something that some friends and I use online. But I have been following the advancements in AI and noticed that you can now use it to describe how you want your app to work in plain language, and AI will code it for you. I cannot write code. I cannot even begin to imagine how to write code. But I have been working with AI for a couple of weeks now and we are making progress. And in a few more weeks, I hope to have a working application. I will get into more details about how fascinating this process has been. But what I want to say here is that AI is giving this non-coder an opportunity to do something he could never do before. I will never replace a real, experienced, imaginative coder. And that's the similarity I see to the writing world. AI will hopefully never replace experienced and imaginative writers. It will give people who really can't write very well the opportunity to do something they can be, frankly, proud of. Believe me, if I can get this app to work I will be as proud of the effort as I have ever been with any of my writing. What does all this mean? I am thinking through the implications.
Deb, AI can be a time-saver. EX: My historical essay on the New England Vampire Panic was approx 1,500 words. After I created a "vampire panic" video, I asked AI to condense my own essay to 300 words to use for my YouTube description. This task would have -- in the past -- taken me no less than an hour. I am happy to assign all my scut work to an uncomplaining AI assistant.
From an ethical standpoint, that's a good observation, but I personally wouldn't trust AI to make editorial recommendations unless I was just looking for obvious, glaring plot holes. I'd much prefer a human beta reader (who I also might ignore). The way I see it, I don't think Dickens or Dostoevsky or the lesser writers of their eras had beta readers. I've been reading literature since I was in third grade. I trust myself a lot more than I trust a lot of others, human or otherwise. This is likely a flaw, but it's my flaw.
I think using it for research is fine. I sometimes use it as a launching pad, then follow up with a heavy duty validation on my own. Sort of like checking the work of someone else's research. But it saves a lot of time, especially if there's a ton of research involved.
I love free A.I. help about what SEO words to include in my press release, summary, etc. 🥩 🥓 Additionally, I am doing far less harm to the world than people who eat meat or pollute the oceans.
Seems like there is something else the article merely alludes to. That is these data centers. I am one of the ones calling for regulation, not because I dislike AI, because I find it quite useful. Like you, I have no need of it in drafting or editing anything. I have decades of experience in creative and technical writing, and in generating compelling visual presentations. AI just isn't really smart yet, but it is fine as an aid to research and to generate some images. If you want labels in the images you really have to work with it so they aren't garbled.
But here is the regulatory issue: It is not just creative use that ought to be regulated, but data centers themselves. Perhaps it is a separate issue. If you think so, I get it. But consider the fights against the construction of new data centers. The problem in the US with that is that we have long been wedged apart by either/or logical fallacies. For example, big oil has told us for years we can either have good jobs OR clean water and air, but not both.
Ah, but why not both? A data center will consume upwards of five million gallons of water a day because massive amounts of water are required for cooling. Most data centers lose 70-80% of the water used in cooling. Some are constructed with better science in mind, and use closed loop cooling systems and recycle wastewater. Others, located in colder climes, can supplement water with air cooling.
What is happening that requires regulation is procedural. For decades in the US, this false choice between jobs and the environment has created a whole lot of ignorance. So, we have local governments who are okaying the construction of data centers without any public meetings, without having anyone else at the table to ensure the incoming centers will address water consumption AND electricity consumption. A proposed data center in Box Elder County, UT, for example, was going to over double the electricity consumption in the entire state! Seriously.
And, silly people, they were concerned about the depletion of their Salt Lake watershed over time and that their electrical bills would increase much faster than inflation. So there was a massive public outrcy and the developer had to back off. They didn't have the right people at the table, you see, and made the ill conceived choice to merely approve the new center administratively.
This also begs another question, Joe, and this is why we need all these new data centers. What are they really for? Why do we have this feverish need to build, build, build? What state are we hoping to achieve with this proliferation of data centers? Because they arguably are parasitical in terms of water and power usage UNLESS the approach to their location and building can be reasonably regulated to ensure closed loop cooling systems, recycling of wastewater, and leveraging solar/wind power to offset electricity consumption. There is even talk of mini fusion reactors to power centers, and those are expected to be grid-ready by 2030.
The bottom line is that progress is what it is. We cannot stop it, but we can ask why, demand purpose (if $X is invested in a new data center, our water supply will do Y, and the electricity consumption will go to Z, offset by A and B renewable energy sources). The problem with the United States is that we have become so very adversarial about everything.
But the question remains - Why do we need all these new data centers? How do they fit into overall sustainable economic development, urban planning and workforce development?
You make an excellent point. I did, in fact, gloss over the issue of data centers. It's kind of an out of sight, out of mind thing. There are no data centers where I live. Like so many other negative aspects of our society, they are relegated to areas in which the people have little power to fight them. But their effect on the environment is already profound, and probably will continue to get worse until more responsible aspects of science catch up. It would be great if the profit oriented segments of our society could work in tandem with the environmentally conscious ones, but we all know it doesn't work that way. The best we can hope for, since as you mentioned we are not going to stop AI progress, is that we can catch up quickly. A significant part of that is educating the population about the pros and cons of AI and the effect that progress is having on the environment. I've noticed that there really isn't much in major media about data centers.
RE Data centers: Read Will Lockett or Cory Doctorow (he's the enshittification guy) on the AI bubble. When it pops, it's going to leave a lot of empty data centers behind in its wake.
Pulitzer Prize winning novelist Colson Whitehead wrote, “People are usually surprised when I admit that I love AI. I use AI for everything, 24/7. I can’t get enough, me and my whole family.” He just doesn’t use it for his writing. He hastened to add, “Except art. If you use it for your art, you’re a freakin’ hack.”
That's hilarious
I appreciate the points argued in this article. It prompts me to reflect on my negative response to last post’s AI image. I didn’t try to rationalise it at the time, but I did choose to share it. As a caveat, I’m not American.
As AI slop increases, I understand there to be a greater role for gatekeepers or curators. There are going to be decent things produced with the help of AI and a lot a lot a lot of subpar things. In the case of the image, perhaps it was that you had selected it and it felt so devoid of effort, the kind of clearly low quality AI aesthetic that makes up so much slop.
Interestingly, before AI could produce images even this well, I enjoyed the somewhat psychedelic images that a data scientist produced using, I think, GANs, some time in 2019ish. Again there was a level of curation though, as he had chosen only his favourites to share, printing and handing them out to us on cards.
You mentioned that it will be a greater role for gatekeepers and curators. Quite a few columnists (some mentioned in the post) have noted that the gap between people who know how to use AI and those who don't is just going to keep getting bigger. Let's hope that those gatekeepers and curators are among those who have the knowledge and experience to use AI for its best purposes.
A.I. helps save my sanity and conserve my energy for "paying markets" and their requirements. WHY? Because there is an entire eco-system of whimsical lit-mags who do NOT pay but who are still determined to suck the marrow from a writer's bones. EX: the lit-mag editor who accepts my CNF and then requests an essay on my inspiration. Or the lit-mag editor who accepts a poem and then demands an "author's note." Blah-blah-blah. Now I grind my WORD doc into A.I. + instructions & it spits out my 200-word "author's note." . . . . . . . . . . . . This gives me more time to devote to zines that actually pay poets and authors. ✒️ 🙏 ‼️ Shame on anyone who exploits another writer by denying payment.
Excellent point. I suspect the lit mag editors who do these kinds of things oh sincere, but it also shows how naïve, and frankly, self-centered they are. They think they are serving the "art," but they fail to see the bigger picture.
Lit-mags and zines that compensate poets and writers deserve my best. Those who do not pay and impose extra drudgery.... I give that to my "assistant" now. LOL