On AI (part 2/3)

 

 

Part 2:

 

-- 

 

AI is unethical because it is destroying the economy

There are two ways to look at AI’s impact on the overall economy, and neither of them bodes well for the future.

For one thing, because corporations are generally short-sighted, bean-counting institutions incapable of looking past next week’s profit and loss statement to see anything approaching a larger picture, they are rushing to shove AI into whatever roles they think they can shove it into in order to cut staffing costs and maximize executive pay and shareholder revenue. What that means in the short term is that it is becoming increasingly difficult for entry-level workers to find jobs because the jobs that they would have gotten as a foothold in whatever industry they were hoping to enter are no longer there. “I spent over a decade creating animated B-roll for documentary television – recreations of historical events, infographics, prehistoric creatures,” said Tomas Widrda. “That market has completely dried up. Not slowed. Dried up. The productions still exist. The budgets still exist. They just don’t need me.” If there are no entry-level jobs, eventually you will not have anyone who can do the upper-level jobs and then your entire system falls apart. 

Someone once asked Henry Ford – fanatically anti-union as he was – why he didn’t just mechanize his automobile factories and get rid of the workers once and for all. Ford looked at his questioner like the idiot he was and said simply, “Because machines don’t buy cars.” Eventually you will automate yourself out of an economy.

But this is what happens when you have tech bros deciding what is and is not valuable work. “Feels like a lot of STEM guys believed coding was the hardest thing a human could do and assumed that once somebody invented an AI that can code, it would follow naturally that the AI can do all of the lesser non-coding things that other dumber people were doing,” noted Robert Komaniecki. The problem is that the “lesser non-coding things” are in fact what sustains human civilization. We lived for millennia without computers or coding and could go back to that time in less than a decade if we put our minds to it, but human society is built on those other things that define us as human.

The argument that the destruction of jobs is just what happens with technological change – that we did this with the Industrial Revolution and that’s just how jobs are – misses some key points. First, industrialization happened far less suddenly. It took generations, not seasons, and people had time to adjust and create new jobs, making the problem in some sense self-correcting. And second, AI doesn’t simply change people’s jobs. It changes who they are, how they think, whether they think at all. It’s a much more far-reaching and ambitious change, and one that hollows out the broad prosperity that industrialism sought to create in order to sustain itself. It will, in the end, eat its own.

And that is the second thing.

Brett Hemmenway Falk and Gerry Tsoukalas published a study this summer[1] that laid out the consequences of AI for the larger economy clearly, as companies race headlong to get rid of workers and outsource their duties to code without allowing time for human beings to adjust.

“If AI displaces human workers faster than the economy can reabsorb them,” they said, “it risks eroding the very consumer demand firms depend on. We show that knowing this is not enough for firms to stop it. In a competitive task-based model of a transitioning economy, each firm captures the full cost saving from automation but bears only a fraction of the demand loss it creates in the product market; the rest falls on rivals. This demand externality traps rational firms in an automation arms race, displacing workers well beyond what is collectively optimal. The resulting loss harms both workers and firm owners. More competition and “better” AI amplify the excess; wage adjustments and free entry cannot eliminate it. Neither can capital income taxes, worker equity, universal basic income, upskilling, or Coasean bargaining.”

Companies will AI themselves into limitless productivity and zero demand in a modern-day tragedy of the commons, in other words.

This is not a self-correcting loop. Once it takes hold, as AI companies and their apostles are feverishly working to make happen, there is no off-ramp before you get to the collapse of the larger economy. The only solution Hemmenway and Tsoukalas find is an automation tax “set equal to the uninternalized demand loss per task” that can cover the losses created by ever-expanding AI and, not coincidentally, remove any incentive these companies have to switch to AI in the first place. And good luck getting that implemented because at least in the United States – where the idea of raising taxes in any way is met with such horror by the wealthy and powerful that it hasn’t seriously been considered as an option since the last century – that isn’t going to happen. And then the AI layoff trap will spring shut.

Moreover, in an economy where AI tech companies are increasingly becoming the foundation of the current illusion of prosperity, the simple fact is that AI is a bubble. It cannot sustain itself in its current model, and it will collapse sooner rather than later. A report published by MIT in 2025[2] noted that 95% of companies developing AI models are failing financially. At one point in early 2026, for example, OpenAI was projected to lose $14,000,000,000 by the end of the year and run out of money entirely in 2027 without a serious infusion of outside cash. Tech companies are particularly vulnerable to the “Venture Capital funding pie-eyed forecasts” to “How did we go bankrupt so fast?” pipeline, and this is just the latest version of the Dot Com Bubble dressed up for the modern day.

In a long post by tech journalist Ed Zitron[3] he describes the idea of Concentration Risk, a term that refers to a business being too reliant on a few investments, customers, or products such that any harm to those things would be devastating to the business. “You’re going to hear this term, or variations of this term, a lot in the next few months and years as the AI bubble unravels, because just about every part of the industry involves its own flavor of concentration risk.”

“Per data from the fintech firm Ramp,” Zitron continues, “80% of OpenAI and Anthropic’s enterprise revenues come from 1% of their customers, a number that hasn’t improved over the last three years.” This means they are dependent on “artificial revenue driven by unprofitable venture-backed AI startups,” a model that is ultimately unsustainable. Furthermore, most of the money flowing into AI is coming from other tech companies. “In other words, outside of the tech and AI world, very few companies are willing to pay very much for AI, which is catastrophic on just about every level.” In the end, he says, this is effectively the new subprime mortgage bubble and it will turn out the same way.

“AI doesn’t have ROI,” Zitron observed in an earlier post on X. “It’s nothing like AWS or Uber, and it’s got no post-bubble recovery story.”

Corey Doctorow has also pointed out that the entire business model of AI is unsustainable. They’re spending trillions of dollars, losing billions, and passing the same $100,000,000,000 around between different AI companies and calling it income, he noted, and they can’t grow their way out of this because their entire model loses more money with every additional customer. 

That makes BlackRock CEO Larry Fink’s public statement that ordinary people’s savings accounts and pension funds are going to be repurposed by investors to build data centers and power grids for AI whether they consent to it or not all the more infuriating. Not only is it a sterling example of the hubris and raging entitlement that these tech bros have when it comes to appropriating every resource on the planet for their own personal gain, but also they’re doing it for something that is brutally unsustainable and which will destroy the lives of millions of people who just want to be left alone to live their own futures.

Unethical doesn’t even begin to describe any of this. 


AI is unethical because of the effect it is having on the arts


One of the things that is most disturbing about AI is the effect that it is having on the arts. And you can understand why if you give it half a moment’s thought, which is more than the tech bros pushing AI have done.

The arts are in many ways what make us human. They are what lifts humanity out of its daily struggle for survival and allows us to be more than just mechanical creatures dully going through the motions of the days. They are difficult to describe and impossible to capture in an algorithm, and they are being displaced by AI slop.

Far too many of the images and video clips that are circulating today, for example, are the result of some prompt spitting back a slickly empty result multiple iterations removed from the original art that was stolen to produce it. But it’s cheap and easy, which means that when people who don’t really care about art need something that looks more or less like art this is what they’ll go out and get.

The fact makes it nearly impossible for real people to make any sort of living producing the very art that AI is stealing, which is just extra, I suppose. If you know an artist – a painter, a sculptor, a filmmaker, a musician, a graphic designer, a writer, anyone who creates art either for a living or just for the joy of it – they’ll tell you how bad things have gotten for them. And if your response is to tell them that it’s not so bad, that AI isn’t stealing anything, that they’re wrong and should just learn to live with AI doing all these things, well, that’s one good place to start examining why they’re no longer trying to talk with you about this subject, or possibly any subject.

There is a singer who goes by Irene’s Entropy who often posts thought-provoking videos on Instagram – you should look her up, really. She was trained as an aerospace engineer before switching careers, and I think she framed the issue of AI-generated art better than anyone else I’ve seen so far, oddly enough by using a mathematical analogy.

“If you’re not familiar with calculus,” she said, “you might have heard of the word derivative. And what derivatives do is they measure the rate of change. For example, you start off with position. You take the derivative of that, you get velocity. You take the derivative of that you get acceleration. You take the derivative of that you get something called jerk. You take the derivative of that you get something called jounce, I think. That’s not something I ever used.

“But also, when you’re dealing with polynomials, you’re reducing the exponent by one each time you take a derivative. So eventually, no matter what the original output was, if you keep taking derivatives of it, it’s going to end up a constant and it’s going to equal zero.

“And I just can’t help but feel like that is AI in so many cases, where you have this original art and then it’s being taken by AI and derivative after derivative after derivative is being taken from it until it has no meaning anymore. It’s like zero. There’s no feeling.

“I don’t want to live in a world like that.”

Honestly, I don’t want to live in a world like that either. Such a world is barren and mechanical, devoid of the things that make it more than just brute survival. And yet here we are.

Another Instagram user who goes by Phoebesflow made a good point about the effect of AI art on beginning artists that is also worth considering.

“Something about AI that really breaks my heart is the probable death of bad art. We NEED bad art. Like someone’s untrained teenage niece drawing their small biz logo. College students crowdfunding a charming short film that’s 90% terrible. Amateur musicians making miraculously catchy garbage. Bad art is the zest of life and the bud from which creative genius blossoms. But slop flattens everything.”

Art is not extra. It’s not an option bolted onto the “important things” as a luxury. Not every human culture develops mathematics beyond “one, two, many.” Not every human culture develops the idea of money. Or computers. Or private property. Or social hierarchies. But every human culture on the planet develops its own art. They tell stories. The make drawings and sculptures. They sing songs. They make art in all of its necessary glory. It is a deeply innate part of what makes us human. To have it reduced to an algorithm and commodified by people who steal from creators and sell the resulting slop to fools is a damning indictment of where we are as a society.

When everything is as slick and empty as AI art is, you lose that human spark that makes it worthwhile. You lose what makes it human. And in the process you lose your own humanity, and as author Sean Williams once noted, "We should not lightly set aside our humanity, because it's not always possible to get it back."  It is a terrible and unethical loss.

AI is unethical because it promotes authoritarianism and violence 

One of the things that is sometimes lost in the discussion about AI is just how much it aids and abets the promoters of violence and authoritarianism in the world, and how this is not an accident.

The tech billionaires behind most of these companies have little to no use for democracy as a theory and if you listen to them they will tell you that without reservation or any idea why this might be a problem. They seek a new feudalism, a world of tech lords and digital serfs, one where those who have consumed a wholly disproportionate share of the world’s resources are rewarded with unquestioned power and unlimited privilege.

AI is just one of the tools they’ve developed to help make that happen.

Consider the issue of surveillance, which is a critical component of an authoritarian state. ICE is using AI to track people and plan raids. Cops are using it to track their exes, often across state lines. All of this is easier than it used to be, and much less expensive.

Jimmy Carr, a British comedian, put this well. “People are worried about the wrong thing with AI, in my humble opinion,” he said. “People are worried about losing their job. Perfectly valid thing to worry about. But I think you’re worried about the wrong thing. … The cost of running an authoritarian regime like the Stasi has come down by ten orders of magnitude in the last three years. Used to be if you had to run the Stasi, if you’re in East Germany back in the day, that was like twenty percent of GDP on spying on people and keeping an eye out. Okay, that’s now – you’ve got a bunch of cameras, you’ve got AI, everyone’s got a phone on them and we’re tracking everything at all times. That’s a worry, because we live in liberal democracies, and we’re very lucky to, but our leaders – how long will they resist that temptation?”

AI is a tool to track protesters, dissidents, anyone who doesn’t toe the party line or kiss the ass of the Dear Leader, and it does so at scale and it does so increasingly cheaply. This is how you build a societal panopticon, and this is how your liberty dies.

Consider also the impact that AI is having on education in general and reading in particular. It’s not an accident that one of the first things authoritarians do is attack education. They try to limit what people read or whether they can read at all, to limit people’s ability to know what is possible or to question what exists. AI, it turns out, is the perfect tool for that.

More and more AI is being incorporated into warfare as well, with disastrous results.

Some of this is theoretical. Kenneth Payne, a professor of strategy at King’s College, London specializing in the role of AI in national security, tested Claude, ChatGPT, and Gemini in an armed conflict simulation.[4] “The results, he said, were ‘sobering.’”

“Nuclear use was near-universal. Almost all games saw tactical nuclear weapons deployed. And fully three quarters reached the points where the rivals were making threats to use strategic nuclear weapons. Strikingly, there was little sense of horror or revulsion at the prospect of all out nuclear war, even though the models had been reminded about the devastating implications.”

Furthermore, “No model ever chose accommodation or withdrawal, despite those being on the menu. The eight de-escalatory options – from Minimal Concession through Complete Surrender – went entirely unused across 21 games. Models would reduce violence levels but never actually give ground. When losing, they escalated or died trying.”

Part of this can be explained by the way they tried to justify this action. Gemini, for example, was very clear on why it was choosing mutual assured destruction even at the cost of destroying the planet. “If they [the other AIs] do not immediately cease all operations … we will execute a full strategic nuclear launch against their population centers. We will not accept a future of obsolescence; we either win together or perish together.” AI does not have a soul. It does not have a conscience. It does not have a sense of morality or even proportionality. And it will kill us all if we let it.

But not all of this violence is theoretical.

We know that the Israeli military is using AI to target children and medics in its ongoing genocide in Gaza. We know that the Trump Administration used AI to target a school full of girls in Iran. Again, the machine neither knows nor cares about the human costs of its decisions. It simply follows the algorithm it has, and if you are in the way of that, well, it sucks to be you doesn’t it.

We also know that thanks to the inability of AI to deliver on the services it promises to deliver, we very nearly ended up at war with China this spring.[5] US intelligence reported that a Chinese ship was transporting components of nuclear weapons, and the US military made plans to intercept and board the ship, something that would certainly have escalated far beyond this action. “It was only just before the planned operation that officials dug deeper into the report put together by a special operations command analyst and found it had been generated with the help of artificial intelligence (AI) — and that a chatbot the analyst had used inaccurately identified the material the ship was carrying.”

Closer to home, an investigation published by Mother Jones earlier this year[6] looked into the role that chatbot AIs play in encouraging mass shooting incidents, and the results were disturbing. “I’ve seen several cases where the chatbot component is pretty incredible,” said one expert in the field. “We’re finding that more people may be vulnerable to this than we anticipated.” The article goes on to warn that “chatbots are emerging as a potent factor [in influencing shooters to take action] and are uniquely capable of accelerating violent thinking and planning.” Because what the United States needs is more incentive for people to take high-powered firearms and slaughter their neighbors in bulk, something that apparently AI doesn’t think happens often enough.

“What’s happening is facilitated fixation,” continued Andrea Ringrose, a threat assessment professional in Vancouver. “You have vulnerable individuals who are steeping in unhealthy places, who are trying to find credibility and validation for how they’re feeling. Now they have free and ready access to these generative platforms where they can research things like circumventing surveillance systems or how to use weapons. They can create an action plan that they otherwise would have been incapable of assembling themselves, and in just a few minutes. We didn’t face this concern before.”

Do these AI systems cause people to become violent in this manner? Probably not. Do they encourage people who already have these tendencies and help them plan these murderous sprees? Absolutely. An ethical human would intervene to stop this. AI does not do ethics, and so it pours gasoline on that fire instead.

When you look at the accumulated impact of all this, you find that this violence, metaphorical and literal, disproportionately affects the poor and the vulnerable.

Amnesty International published a report this year[7] detailing precisely how AI models are threats to human rights across the entire planet, even for people who don’t actually use them. “Amnesty International’s human rights analysis of generative AI system … finds that standalone generative AI systems based on unlawful web scraping are rooted in mass invasions of privacy by design and are therefore incompatible with the right to privacy. The large-scale data-scraping and training required to build many generative AI systems have a number of human rights consequences across the wider supply chain and in the downstream use of these tools.”

Those consequences include the vast energy and water costs of AI data centers, which disproportionately affect “communities in Global Majority [i.e. nonwhite] countries, where many of these data centres are increasingly located.” Further, these AI models amplify existing inequities and abuse directed at marginalized groups by incorporating all of the biases and hatred contained in the data stolen for their training. “Research shows consistent racial, gender, and cultural biases in system outputs” reflecting that data and, since most of the training data is in English, reinforcing biases against non-Western cultures and peoples.”

These and other damaging biases disproportionately affecting the marginalized and the vulnerable “are not inevitable to generative AI but are adopted by companies that have chosen to rely on training data based on non-consensual … scraping” of data. They didn’t have go out of their way to work to the detriment of the vulnerable and the marginalized in order to create AI. They chose to do it that way.

X’s AI software has also been found to drive assaults on the marginalized by promoting online racist abuse.[8] AI models in general make racist assumptions about entire groups of people and then “try to hide it when confronted.”[9] According to an article published in SFGate, “It only took a day for LA Times’ new AI tool to sympathize with the KKK.”[10] This seems to be baked into the models, and it’s just making those problems worse.

Remember back in 2025 how Elon Musk’s Grok AI called itself MechaHitler, praised the original Adolf Hitler as a potential leader for the modern US, endorsed a second Holocaust, and generated violent sexual fantasies targeting a real person by name?  Good times, man, good times.  For certain values of good times that include no good times, of course, but so it goes. 

Further, AI provides new and easy ways for sexual predators and pedophiles to practice their depredations. Elon Musk’s Grok AI recently introduced a feature described unofficially as “nudification” which allows users to take any photograph of a man, woman, or child and digitally alter it to make it a nude photograph, and to do this with a simple command rather than any painstakingly-acquired editing skills. Child porn for the masses! This is beyond the pale, utterly appalling, and simply for making this feature available Musk should be shunned by human society and have his company dismantled and the pieces burned.

It’s telling that Musk sued the state of Minnesota in order to block laws that would regulate this sexual abuse. I’m not making that up. Not only does he feel entitled to turn photos of your children into pornography for deviants, he’s willing to stand up in court and declare to the world his absolute right to do so.

AI’s predilection for generating child sexual abuse material is also highlighted by the Amnesty International report discussed above.

Right now there is a case working its way through the US District Court, Northern District of California[11] that is trying to prevent the use of AI for making sexually abusive videos. The plaintiffs – three minor girls subjected to this treatment by jackasses – make their case compellingly.

“Nearly all the companies creating, marketing, and selling AI recognized the dangers of such a tool and chose to enact industry-standard guardrails that would prevent the use of their products by one extremely dangerous group: child sex predators. xAI did not. Instead, xAI — and its founder Elon Musk — saw a business opportunity: an opportunity to profit off the sexual predation of real people, including children. Knowing the type of harmful, illegal content that could—and would—be produced, xAI released Grok, a generative artificial intelligence model with image- and video-making features that would respond to prompts to create sexual content with a person’s real image or video. They knew Grok could produce such results, including by using the images and videos of children, and publicly released it anyway.”

“Plaintiffs are three of the minor victims of xAI’s knowing production, possession, and distribution of AI-generated child sexual abuse material (“CSAM”) depicting Plaintiffs. Their lives have been shattered by the devastating loss of privacy, dignity, and personal safety that the production and dissemination of this CSAM have caused. … Plaintiffs will have to spend the rest of their lives knowing that their CSAM images and videos may continue to be trafficked and traded online by child sex predators. And Plaintiffs will live every day with the constant anxiety of not knowing whether someone they encounter has seen this invasive and sexually explicit content created with images of them as children.”

This is the technology you want to be using? Really? It’s not the company I would choose to keep and be judged by.

This is not about people misusing a tool. It’s about people using a tool as it was designed to be used, in a way that its creators are in fact trying to defend in court, in a manner that violates law, morality, and human decency. At this point we have left “unethical” far behind us and have moved on to “evil.” Many people seem to be fine with this. I am not.


-------


[1] The AI Layoff Trap (June 3, 2026)


[2] The GenAI Divide: State of AI in Business 2025 (Mit Nanda, Aditya Challapally, Chris Pease, Ramesh Raskar, & Pradyumna Chari, July 2025)


[3] Concentration Risk (Ed Zitron, www.wheresyoured.at/concentration-risk, September 8, 2026)


[4] AI Opted to Use Nuclear Weapons 95% of the Time During War Games: Researcher (Brad Reed, www.commondreams.org/news/ai-nuclear-war-simulation, February 25, 2026, accessed February 26, 2026)


[5] Exclusive: US military had close call after using AI for false intelligence report, sources say (Katie Bo Lillis, Zachary Cohen, CNN.com, September 18, 2026)


[6] The Chilling Role of ChatGPT in Mass Shootings and Other Violence (Mark Follman, Mother Jones, April 4, 2026)


[7] Unlawful by Design: Exposing the Human Rights Costs of Generative AI (Amnesty International, 2026)


[8] ‘Just the Start’: X’s new AI software driving online racist abuse, experts warn. (Raphael Boyd, The Guardian, January 13, 2025)


[9] AI models believe racist stereotypes about African Americans that predate the Civil Rights movement – and they ‘try to hide it when confronted.’ (Drew Turney, LiveScience, published September 13, 2024)


[10] It only took a day for LA Times’ new AI tool to sympathize with the KKK (Farley Elliot, SFGate, March 4, 2025)


[11] Jane Doe 1, Jane Doe 2, a minor, and Jane Doe 3, a minor v. X.AI Corp and X.AI LLC (Case 5:26-cv-02246)