With solidarity, legislation, class action lawsuits, and a little help from Guillermo del Toro. Plus, the Luddite Renaissance continues, and I open the mailbag.
Isn’t this AI hype JUST like the OUTSOURCING craze? You couldn’t get VC funding if you didn’t promise to outsource your labor. It seemed like all call centers moved to India. It all failed because dirt cheap labor soon got COMPETITIVE!
I can’t believe the upfront investment in millions they are paying out for unproven economic gains in Labor.
So for all the people out there that desire to avoid AI at all costs what exactly is the end game here? Because right now the entire anti-AI enterprise operates as if it is born out of principle or some ethical system and I must say it seems largely, irrational, illogical and driven by fear. What is hoped to be achieved? Given that AI and ML is in everything, nearly all our technologies, how is an anti-AI purist to avoid non Slop AI like;
I can't tell if you're being willfully obtuse here or if you legitimately believe if artists and workers are opposing AI in any capacity they must oppose Autocorrect as well. This is obviously not the case. So in the interest in having good faith discussion here, I'll just say that the artists are opposed to commercial generative AI technologies being used and sold to automate their work; no one is demanding an end to email filters.
(It is also perhaps worth noting that "AI" in 2025 is cultural shorthand for gen AI and products sold by OpenAI and big tech, and does not typically indicate machine learning techniques deployed by researchers etc, and almost everyone who follows this space recognizes that.)
The idea that we need to accept AI in every formation or none at all is pretty profoundly silly. It is, for instance, quite possible to say 'Autocorrect is good for society while large language models used in contexts that erode artists' working conditions are not'.
So why won’t modern artists and workers leverage these new AI tools to improve their craft and how they work? Historically artists have always leveraged new technical tools to improve and expand their craft, from painting to photography to digital art. Take for instance the use of the Camera Lucida, the Camera Obscura and other optical devices, to aid in the drawing process used by the old masters. Perhaps the art world should focus more on leveraging technology and what its existence offers like their forebears did instead of arguing against its use.
Old Masters use of optical devices to aid drawing and painting camera obscura lucida
None of these technologies was built on the scraped labor of the very artists you're suggesting use them. None of these required underpaid laborers in the global South to assess graphic images of rape and other violence to clean up outputs, often resulting in PTSD. None of these is built on datasets that include child sexual abuse material. And porn, too...
"Later, during the development of DALL-E 3 [OpenAI’s image generator] , when the data imperative had grown even larger, the research team decided that sexual images were no longer just a ‘nice to have’ but a ‘need to have.’ The share of pornographic images on the internet was so large that removing them shrank the training dataset enough to notably degrade the model’s performance. In particular, it made the model worse at generating faces of women and people of color due to the discovery that Deborah Raji had made as a Clarifai intern: A significant share of the online content depicting both groups is sexually explicit. For the same reasons, the researchers left in some other kinds of disturbing images.” (238) Karen Hao, Empire of AI.
There is some discussion here on substack of artists who have tested to see where in the process they might interject AI to e.g. clean up their sketches.
IMO for quick output, sure, gains are made but I’m not so sure the artist themselves learn and GROW from this which is the essence of art.
I feel like your overall points hinge on strawmaining & downsizing AI's current state to assume that people against it just don't like change which is far from what anyone has actually said. Saying that people can't be against AI because they used it is like saying that you can't be against cars because you rely on any form of transportation. At best this feels misinfomed about why people are against it & at worst very dishonest rhetoric that follows the "you can't be against captialism becuase you live in it" line of logic.
Artists and by extension people at the end of the day want tech companies to at very least do the bare minimal when it comes to not being reckless or actively trying to hunt & hurt their livelihood.
The fact you tried to compared AI ot most of these devices you linked really highlights you don't why people don't like the current state of AI.
FWIW, my take is not that this readership, or the author, wish to "avoid AI at all costs", but rather to continue to explore and promote ways to avoid or mitigate the harm AI is causing many for the substantial benefit of a few already very wealthy individuals. Yes, we all benefit in ways you have outlined (though I'm not sure how many things on your list are AI, or just huge computing power search engines - you may be better qualified than I to answer that), but I would forgo most of those benefits if it meant avoidance of large scare job losses.
Then by implication so is linear regression, which was introduced in the late 19th century. That's the starting point of books like Hastie, Tibshirani, and Friedman's "The elements of statistical learning". Most of what's commonly known as machine learning could fairly be called glorified regression, albeit usually not linear. Even artificial neural networks could fairly be called that; as Hastie et al. remark (p. 392):
"There has been a great deal of hype surrounding neural networks, making them seem magical and mysterious. As we make clear in this section, they are just nonlinear statistical models, much like the projection pursuit regression model discussed above."
However, "AI" isn't just statistical models but a set of claims that have rarely, if ever, been made about what's commonly known as machine learning: that it is or will soon be "intelligent" in some broad sense; that it could or should take over most intellectual labor from humans; that it deserves large shares of the world's investment capital and electricity generation; and so on. From its origin in the 1950s, "AI" has been at least as much a marketing term as a technical one, and never more so than at present. Moreover, most machine-learning systems aren't trained on data appropriated from humans without their approval or compensation. (The most conspicuous exceptions, which, however, don't constitute a majority of machine-learning systems, belong to exploitising platforms like Google and Facebook. I doubt most readers of *Blood in the machine* like them any better than they like "AI".)
Asserting that machine learning is just a subset of "AI" or that "AI" is just an elaboration of machine learning amounts to disregarding those claims and thefts. Intentionally or not, it's misleading. Technology isn't just technique.
By the way, Google-Scholar me if you care. Among other things, you'll find I'm a coauthor of an article applying a support vector machine to analyzing gene expression. As a "data scientist" at a company, I often used machine learning (e.g., for classifying reviews supposedly by our users, but some of them were by bots). I have advanced degrees in biology, math, and physics from name-brand universities. I also worked as a researcher in the computer science department at UC Berkeley and at the Swedish Institute of Computer Science (where the people who worked on what some would call "AI" tended to cringe at the term). So I'm no technophobe or technical illiterate.
If anyone's interested in submitting to LOOMING issue 2, here's a link to the call:
https://drive.proton.me/urls/DTXT46463R#GPNQIUZU1z3m
Glad to see that my fellow artists are still fighting back and that your work is making the rounds! Also, go Jays!
Isn’t this AI hype JUST like the OUTSOURCING craze? You couldn’t get VC funding if you didn’t promise to outsource your labor. It seemed like all call centers moved to India. It all failed because dirt cheap labor soon got COMPETITIVE!
I can’t believe the upfront investment in millions they are paying out for unproven economic gains in Labor.
That is the talk in the air since late 2022. That VCs revoked seed investment from startups that were not working on AI.
Brian, I love BLOOD IN THE MACHINE! I agree 100% with Reid Southen about AI. A cartoon I did on this abomination: https://open.substack.com/pub/annecantstandit/p/ai-aye-yai-yai?r=qowdg&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false
Thanks for the Looming pdf, and doing the Lord's work.
I think the intellectual property litigation is probably our best bet for resistance.
The cat may be out of the bag but the cat sucks... Not good enough for professional level work, at all.
Was curious if you have thoughts on the book Against The Machine? I've heard good things
So for all the people out there that desire to avoid AI at all costs what exactly is the end game here? Because right now the entire anti-AI enterprise operates as if it is born out of principle or some ethical system and I must say it seems largely, irrational, illogical and driven by fear. What is hoped to be achieved? Given that AI and ML is in everything, nearly all our technologies, how is an anti-AI purist to avoid non Slop AI like;
Virtual Assistants
Map and Navigation Apps
Autocorrect
All Search Engines
Facial Recognition technology
Chatbots
Streaming Recommendations
Book Recommendations
Music Recommendations
All Online Shopping
Smart Home Devices
Fraud Detection
Email Filters
Healthcare Diagnostics
Social Media Feeds
Self-Driving Cars
Security Surveillance
Education Platforms
Finance Apps
Weather Forecasting
Smart Appliances
Energy Management
Personalized News Feeds
Wearable Health Devices
Smart Alarm Clocks
Content Creation
Grocery Shopping Systems
On and on …
I can't tell if you're being willfully obtuse here or if you legitimately believe if artists and workers are opposing AI in any capacity they must oppose Autocorrect as well. This is obviously not the case. So in the interest in having good faith discussion here, I'll just say that the artists are opposed to commercial generative AI technologies being used and sold to automate their work; no one is demanding an end to email filters.
(It is also perhaps worth noting that "AI" in 2025 is cultural shorthand for gen AI and products sold by OpenAI and big tech, and does not typically indicate machine learning techniques deployed by researchers etc, and almost everyone who follows this space recognizes that.)
The idea that we need to accept AI in every formation or none at all is pretty profoundly silly. It is, for instance, quite possible to say 'Autocorrect is good for society while large language models used in contexts that erode artists' working conditions are not'.
So why won’t modern artists and workers leverage these new AI tools to improve their craft and how they work? Historically artists have always leveraged new technical tools to improve and expand their craft, from painting to photography to digital art. Take for instance the use of the Camera Lucida, the Camera Obscura and other optical devices, to aid in the drawing process used by the old masters. Perhaps the art world should focus more on leveraging technology and what its existence offers like their forebears did instead of arguing against its use.
Old Masters use of optical devices to aid drawing and painting camera obscura lucida
https://www.youtube.com/watch?v=uq9D0PMGYQc&t=5s
David Hockney, The Lost Secrets of the Old Masters
https://www.youtube.com/watch?v=jMRpmqeKg-g
History of the Camera Lucida Drawing Tool
https://www.youtube.com/watch?v=R_p9z-tRzCs
David Hockney's Secret Knowledge
https://www.youtube.com/watch?v=R-0UXBcjlRY
None of these technologies was built on the scraped labor of the very artists you're suggesting use them. None of these required underpaid laborers in the global South to assess graphic images of rape and other violence to clean up outputs, often resulting in PTSD. None of these is built on datasets that include child sexual abuse material. And porn, too...
"Later, during the development of DALL-E 3 [OpenAI’s image generator] , when the data imperative had grown even larger, the research team decided that sexual images were no longer just a ‘nice to have’ but a ‘need to have.’ The share of pornographic images on the internet was so large that removing them shrank the training dataset enough to notably degrade the model’s performance. In particular, it made the model worse at generating faces of women and people of color due to the discovery that Deborah Raji had made as a Clarifai intern: A significant share of the online content depicting both groups is sexually explicit. For the same reasons, the researchers left in some other kinds of disturbing images.” (238) Karen Hao, Empire of AI.
There is some discussion here on substack of artists who have tested to see where in the process they might interject AI to e.g. clean up their sketches.
IMO for quick output, sure, gains are made but I’m not so sure the artist themselves learn and GROW from this which is the essence of art.
I feel like your overall points hinge on strawmaining & downsizing AI's current state to assume that people against it just don't like change which is far from what anyone has actually said. Saying that people can't be against AI because they used it is like saying that you can't be against cars because you rely on any form of transportation. At best this feels misinfomed about why people are against it & at worst very dishonest rhetoric that follows the "you can't be against captialism becuase you live in it" line of logic.
Artists and by extension people at the end of the day want tech companies to at very least do the bare minimal when it comes to not being reckless or actively trying to hunt & hurt their livelihood.
The fact you tried to compared AI ot most of these devices you linked really highlights you don't why people don't like the current state of AI.
Did an AI make this list for you?
No I just been around a while
FWIW, my take is not that this readership, or the author, wish to "avoid AI at all costs", but rather to continue to explore and promote ways to avoid or mitigate the harm AI is causing many for the substantial benefit of a few already very wealthy individuals. Yes, we all benefit in ways you have outlined (though I'm not sure how many things on your list are AI, or just huge computing power search engines - you may be better qualified than I to answer that), but I would forgo most of those benefits if it meant avoidance of large scare job losses.
ML is a subset of AI
Then by implication so is linear regression, which was introduced in the late 19th century. That's the starting point of books like Hastie, Tibshirani, and Friedman's "The elements of statistical learning". Most of what's commonly known as machine learning could fairly be called glorified regression, albeit usually not linear. Even artificial neural networks could fairly be called that; as Hastie et al. remark (p. 392):
"There has been a great deal of hype surrounding neural networks, making them seem magical and mysterious. As we make clear in this section, they are just nonlinear statistical models, much like the projection pursuit regression model discussed above."
However, "AI" isn't just statistical models but a set of claims that have rarely, if ever, been made about what's commonly known as machine learning: that it is or will soon be "intelligent" in some broad sense; that it could or should take over most intellectual labor from humans; that it deserves large shares of the world's investment capital and electricity generation; and so on. From its origin in the 1950s, "AI" has been at least as much a marketing term as a technical one, and never more so than at present. Moreover, most machine-learning systems aren't trained on data appropriated from humans without their approval or compensation. (The most conspicuous exceptions, which, however, don't constitute a majority of machine-learning systems, belong to exploitising platforms like Google and Facebook. I doubt most readers of *Blood in the machine* like them any better than they like "AI".)
Asserting that machine learning is just a subset of "AI" or that "AI" is just an elaboration of machine learning amounts to disregarding those claims and thefts. Intentionally or not, it's misleading. Technology isn't just technique.
By the way, Google-Scholar me if you care. Among other things, you'll find I'm a coauthor of an article applying a support vector machine to analyzing gene expression. As a "data scientist" at a company, I often used machine learning (e.g., for classifying reviews supposedly by our users, but some of them were by bots). I have advanced degrees in biology, math, and physics from name-brand universities. I also worked as a researcher in the computer science department at UC Berkeley and at the Swedish Institute of Computer Science (where the people who worked on what some would call "AI" tended to cringe at the term). So I'm no technophobe or technical illiterate.