I work as a sales/mktg director at a small university press. We use human artists for covers. If authors come and say they’ve got this AI design idea I shut it down immediately. “We use human artists and we’ll do a great cover for you but we’re not using generative AI.” It’s not much but it’s what I can do to push back. Ditto as a consumer. Big music fan but if I see a band/artist with an AI cover I won’t listen buy or stream!
To one of the artists who commented that it was *demoralizing* because the public cannot tell the difference between good and bad art—well, that’s been the case ever since cave people started outlining the shape of animals in caves.
The insult to injury part of that insight is that the clueless/tasteless people consider the AI slop fascinating and engaging. It’s very much a crow-like response.
On another note, as an artist-writer and producer on Substack for the past five years, I’ll share that whenever I visit a newsletter that is steeped in generative AI images, I instantly recoil.
There’s something off putting and unctuous about its vibe. It damages what might have been my interest in that newsletter’s topics and output. Sad.
As you have pointed out before, the Luddite movement was about the unfair redistribution of income, from artisans to factory owners. Are artists any different from artisanal weavers, trades involving horses, Vaudeville entertainers, etc., etc.? Wasn't the development of machine weaving beneficial to teh populations? Wasn't the move to ICE move power an improvement over horses, if only because engines didn't shit over the roads? Weren't movies, and later TV, beneficial to the population? AI currently produces inferior output in the various arts, but at some point, it will be on a par with human output. It could certainly eliminate awful movies with awful scripts and bad acting today.
The issue is how do these "artisans" transfer to other industries to make a living, especially if the transition is very fast. I don't believe the BS about AI replacing most human jobs, but I do think we have to think about what new jobs will be created/needed, and how artisans will be rewarded for their effort.
Railing against job losses is pushing against the tide. A "Brian Merchant" complaining about the horse trade job losses as the use of "horseless carriages" replaced horse-drawn transport would not have benefitted society if the horse-related industries had succeeded in preventing fossil fuel-powered vehicles from replacing them. I won't say progress, as AI hasn't created new art styles in any art, AFAIK. Do we want to live with redoing old art subjects and styles? This is what happened with painting until the new wave of artists created new styles of art. We will still want these new ideas. Perhaps new ideas can be IP-protected to ensure use of the style compensates the human artist[s].
What the focus needs to be is the creation of new jobs that are related to the industries replacing the artistic artisans to allow them to transition to these new jobs that hopefully will be better remunerated.
The new jobs generative AI has made so far are data workers doing data labeling for pennies and reinforcement with human feedback (RLHF) work on contract for pennies, whether it's in the global south or in the U.S. Time Magazine and Karen Hao have reported on this. She did an interesting interview discussing how white collar work could be transitioning from full time jobs to a precarious gig model.
I would say that cars were more broadly useful than generative AI is and to be frank, given the climate crisis that that form of transportation contributed to, we would have been better off sticking with horses in hindsight. TV began the distortion of our attention spans. My point is, we should ask what the benefits of a technology actually are (if any) and if it's worth what we're giving up. If it's not, we absolutely can unite to resist and refuse it.
We can learn from what happened to industrial workers who lost their livelihoods to offshoring and automation. To paraphrase Timothy Snyder, we have an advantage because we know that history and can organize accordingly.
I would argue that the general problem you address, that displaced workers end up in lower-paying jobs, is not a feature of the technology, but the way we allow the capitalist system to work. This isn't an absolute, but rather an agreed model supported by economists. But clearly other systems do work, even if not so brutally efficient. As we have seen with offshoring, if the supply chain breaks or is deliberately broken, either a new offshore location is needed, or the work returns onshore. (The US tariffs are NOT the way to force this!)
We do have minimum wages to prevent the scenario in Steinbeck's "The Grapes of Wrath" where unscrupulous farmers used desperation to drive down wages. These days, corporations avoid teh minimum wage and required benefits payments by making work in gig work, or creating zero-hours work contracts. All these slimy angles can be eliminated with serious labor laws. We see child labor being reintroduced in some states, something we thought was banned many decades ago.
I suspect you are too young to remember horse-drawn carts. Horses left a mess. In the Victorian era, the smell was awful, and dead horses in teh road added to the mess. Cars pollute, but they are much cleaner to live with. They also don't eat, which requires more land to grow their food - grass for hay, oats, and other vegetables.
As for deciding what technology to adopt and use, well, the US has the Amish community as teh example of how that goes. Do you really want some authority telling you what technology you are allowed to use? We do that by restricting some technologies. OTOH, firearms are very dangerous, yet are given extraordinary leeway to be owned by individuals, with the restrictions on their use pared away by successive legal rulings. Data centers are very energy-intensive today, but we can already run smaller models locally. Technology will make teh models smaller, and teh hardware more efficient, until we will laugh at the clunky size of current AI hardware. The only large datacenters will be used to build models, rather than only hosting the models for inference. Either people will buy the model as we used to do with software, or subscribe to a model delivery service to receive the updated versions. My hope is that specialized models will emerge. Smaller, low-cost/free, and able to run on modest consumer hardware. When the end user can enter the specialty training data and run training models, then we should get a flowering of AI applications, just as software applications are used to create content of all sorts. Hopefully teh data centers can be repurposed, but if not, they will either be torn down or left to ror, creating the new "rust/silicon belt".
I adhere to Kevin Kelly's belief that technology is often a two-edged sword with positives largely balanced with negatives, but that overall there is a small net benefit that accumulates, improving the human world. (Climate change is fixable, but the vested interests have found ways to protect their businesses and stymie the needed change. That is a failing of our political economy, not technology per se.)
I think both the technology and the economic system that birthed it are both problematic. Generative AI is an inherently extractive technology that lends itself to the enclosure of the internet, human knowledge and creativity. That's why Sam Altman can state, in public, that OpenAI wants to sell intelligence like a utility, with everyone having to go to them for it.
I know that the horses were smelly and messy, but they didn't contribute to yearly floods and wildfires. Agree to disagree.
I don't want an authority telling us what technology we're allowed to use; that's in a sense what we have now with this genAI rollout. I want people to be able to decide, through a democratic, grassroots process. What we have now are corporations forcing generative AI on us by putting it in everything and the governments they've promised economic growth and surveillance power to do almost nothing to try and avoid stifling "innovation".
I agree that the current all-purpose models are too big and they will fail. I hope that data centres can be turned into community centres where people can come together.
For me the most brutal ai consequence is the artists that have spent decades learning their craft, which then got fed into the ai so much that now what the artist naturally paints like looks like ai.
This article resonated with me because I've lived through several versions of this story.
As a millwork estimator in the 1980s, I loved the work. I had been a cabinetmaker and shipwright as a teenager, so the transition felt natural and creative. Then estimating software arrived and changed the profession. Later, the internet made it easier to move manufacturing to lower-cost regions, which largely ended the career path I had built. I eventually moved into technical recruiting, only to watch online recruiting platforms reshape that industry as well.
The displacement of work by new technology is an old story that becomes new again with every generation. The initial reaction is often frustration, fear, or anger because something valuable is clearly being lost. Artists today seem to be experiencing that same moment.
What I've learned is that adaptation rarely begins with acceptance. It usually begins with being upset. Over time, some people find ways to adjust, develop new skills, and discover opportunities that weren't visible at the beginning. That process isn't easy, and it isn't fair in every case, but it may be one of the recurring features of technological change.
The difficult question isn't whether technology will continue reshaping creative work. History suggests it will. The question is how individuals and societies help people navigate those transitions while preserving the human skills and experiences that remain valuable.
I work as a sales/mktg director at a small university press. We use human artists for covers. If authors come and say they’ve got this AI design idea I shut it down immediately. “We use human artists and we’ll do a great cover for you but we’re not using generative AI.” It’s not much but it’s what I can do to push back. Ditto as a consumer. Big music fan but if I see a band/artist with an AI cover I won’t listen buy or stream!
To one of the artists who commented that it was *demoralizing* because the public cannot tell the difference between good and bad art—well, that’s been the case ever since cave people started outlining the shape of animals in caves.
The insult to injury part of that insight is that the clueless/tasteless people consider the AI slop fascinating and engaging. It’s very much a crow-like response.
On another note, as an artist-writer and producer on Substack for the past five years, I’ll share that whenever I visit a newsletter that is steeped in generative AI images, I instantly recoil.
There’s something off putting and unctuous about its vibe. It damages what might have been my interest in that newsletter’s topics and output. Sad.
More great coverage from you Brian, thank you.
Interesting to see this all laid out so clearly. Confirms the ‘vibes’, but it’s useful to have it as research.
@Creative PEC (where I day job) had a seminar this week about research at Macquarie University which had looked at author attitudes towards AI (and the prospect of ‘fair’ licensing) which might be of related interest: https://creativepec.substack.com/p/authors-ai-and-the-economics-of-creativity
Thanks Simon, hadn't seen that
They had to call it generative because they couldn't use creative. That is the difference between human and machine.
As you have pointed out before, the Luddite movement was about the unfair redistribution of income, from artisans to factory owners. Are artists any different from artisanal weavers, trades involving horses, Vaudeville entertainers, etc., etc.? Wasn't the development of machine weaving beneficial to teh populations? Wasn't the move to ICE move power an improvement over horses, if only because engines didn't shit over the roads? Weren't movies, and later TV, beneficial to the population? AI currently produces inferior output in the various arts, but at some point, it will be on a par with human output. It could certainly eliminate awful movies with awful scripts and bad acting today.
The issue is how do these "artisans" transfer to other industries to make a living, especially if the transition is very fast. I don't believe the BS about AI replacing most human jobs, but I do think we have to think about what new jobs will be created/needed, and how artisans will be rewarded for their effort.
Railing against job losses is pushing against the tide. A "Brian Merchant" complaining about the horse trade job losses as the use of "horseless carriages" replaced horse-drawn transport would not have benefitted society if the horse-related industries had succeeded in preventing fossil fuel-powered vehicles from replacing them. I won't say progress, as AI hasn't created new art styles in any art, AFAIK. Do we want to live with redoing old art subjects and styles? This is what happened with painting until the new wave of artists created new styles of art. We will still want these new ideas. Perhaps new ideas can be IP-protected to ensure use of the style compensates the human artist[s].
What the focus needs to be is the creation of new jobs that are related to the industries replacing the artistic artisans to allow them to transition to these new jobs that hopefully will be better remunerated.
The new jobs generative AI has made so far are data workers doing data labeling for pennies and reinforcement with human feedback (RLHF) work on contract for pennies, whether it's in the global south or in the U.S. Time Magazine and Karen Hao have reported on this. She did an interesting interview discussing how white collar work could be transitioning from full time jobs to a precarious gig model.
I would say that cars were more broadly useful than generative AI is and to be frank, given the climate crisis that that form of transportation contributed to, we would have been better off sticking with horses in hindsight. TV began the distortion of our attention spans. My point is, we should ask what the benefits of a technology actually are (if any) and if it's worth what we're giving up. If it's not, we absolutely can unite to resist and refuse it.
We can learn from what happened to industrial workers who lost their livelihoods to offshoring and automation. To paraphrase Timothy Snyder, we have an advantage because we know that history and can organize accordingly.
I would argue that the general problem you address, that displaced workers end up in lower-paying jobs, is not a feature of the technology, but the way we allow the capitalist system to work. This isn't an absolute, but rather an agreed model supported by economists. But clearly other systems do work, even if not so brutally efficient. As we have seen with offshoring, if the supply chain breaks or is deliberately broken, either a new offshore location is needed, or the work returns onshore. (The US tariffs are NOT the way to force this!)
We do have minimum wages to prevent the scenario in Steinbeck's "The Grapes of Wrath" where unscrupulous farmers used desperation to drive down wages. These days, corporations avoid teh minimum wage and required benefits payments by making work in gig work, or creating zero-hours work contracts. All these slimy angles can be eliminated with serious labor laws. We see child labor being reintroduced in some states, something we thought was banned many decades ago.
I suspect you are too young to remember horse-drawn carts. Horses left a mess. In the Victorian era, the smell was awful, and dead horses in teh road added to the mess. Cars pollute, but they are much cleaner to live with. They also don't eat, which requires more land to grow their food - grass for hay, oats, and other vegetables.
As for deciding what technology to adopt and use, well, the US has the Amish community as teh example of how that goes. Do you really want some authority telling you what technology you are allowed to use? We do that by restricting some technologies. OTOH, firearms are very dangerous, yet are given extraordinary leeway to be owned by individuals, with the restrictions on their use pared away by successive legal rulings. Data centers are very energy-intensive today, but we can already run smaller models locally. Technology will make teh models smaller, and teh hardware more efficient, until we will laugh at the clunky size of current AI hardware. The only large datacenters will be used to build models, rather than only hosting the models for inference. Either people will buy the model as we used to do with software, or subscribe to a model delivery service to receive the updated versions. My hope is that specialized models will emerge. Smaller, low-cost/free, and able to run on modest consumer hardware. When the end user can enter the specialty training data and run training models, then we should get a flowering of AI applications, just as software applications are used to create content of all sorts. Hopefully teh data centers can be repurposed, but if not, they will either be torn down or left to ror, creating the new "rust/silicon belt".
I adhere to Kevin Kelly's belief that technology is often a two-edged sword with positives largely balanced with negatives, but that overall there is a small net benefit that accumulates, improving the human world. (Climate change is fixable, but the vested interests have found ways to protect their businesses and stymie the needed change. That is a failing of our political economy, not technology per se.)
I think both the technology and the economic system that birthed it are both problematic. Generative AI is an inherently extractive technology that lends itself to the enclosure of the internet, human knowledge and creativity. That's why Sam Altman can state, in public, that OpenAI wants to sell intelligence like a utility, with everyone having to go to them for it.
I know that the horses were smelly and messy, but they didn't contribute to yearly floods and wildfires. Agree to disagree.
I don't want an authority telling us what technology we're allowed to use; that's in a sense what we have now with this genAI rollout. I want people to be able to decide, through a democratic, grassroots process. What we have now are corporations forcing generative AI on us by putting it in everything and the governments they've promised economic growth and surveillance power to do almost nothing to try and avoid stifling "innovation".
I agree that the current all-purpose models are too big and they will fail. I hope that data centres can be turned into community centres where people can come together.
Wow, thank you for all of this!
Brian, thanks so much for coming to OSU. And signing BITM!
Cheers Kevin, the pleasure was all mine!
For me the most brutal ai consequence is the artists that have spent decades learning their craft, which then got fed into the ai so much that now what the artist naturally paints like looks like ai.
This article resonated with me because I've lived through several versions of this story.
As a millwork estimator in the 1980s, I loved the work. I had been a cabinetmaker and shipwright as a teenager, so the transition felt natural and creative. Then estimating software arrived and changed the profession. Later, the internet made it easier to move manufacturing to lower-cost regions, which largely ended the career path I had built. I eventually moved into technical recruiting, only to watch online recruiting platforms reshape that industry as well.
The displacement of work by new technology is an old story that becomes new again with every generation. The initial reaction is often frustration, fear, or anger because something valuable is clearly being lost. Artists today seem to be experiencing that same moment.
What I've learned is that adaptation rarely begins with acceptance. It usually begins with being upset. Over time, some people find ways to adjust, develop new skills, and discover opportunities that weren't visible at the beginning. That process isn't easy, and it isn't fair in every case, but it may be one of the recurring features of technological change.
The difficult question isn't whether technology will continue reshaping creative work. History suggests it will. The question is how individuals and societies help people navigate those transitions while preserving the human skills and experiences that remain valuable.
Thanks for sharing this research, confirms what a lot of us suspected… there’s nothing to like about this technology for creative work.
Honestly, I would like our tech overlords to provide us discarded artists with affordable suicide kits