AI and the erosion of the middle class

My first real adult job wasn't very glamorous. I was an assistant plating manager for a food prep company, and my entire purpose was to make the plating room faster. I communicated with the kitchen to determine what was being made and which components were ready, then delivered them to the platers so they could plate as quickly as possible. I would run around with a printed sheet of all the meals for the day and a pretty good mental map of who needed what, and I was pretty good at it. 

I’ve been thinking about that job a lot lately, in the wake of so much AI progress, because somewhere in the last two years, I've become almost certain that Claude or ChatGPT could do it better than I ever did. I know it sounds harsh, but AI can already replace most of us. I'm not talking about some speculative future, some “maybe one day”, I mean right now. Today. AI could have tracked inventory faster, anticipated lags I would never have seen, and never would have needed a bathroom break or a paycheck. Understand, I’m not trying to be self-deprecating here, but what I would have called cool technological progress a few years ago, I now see as a symptom of a much deeper problem.  

If that’s already true for my plating job, it’s true for so many more. It’s true for possibly more jobs than people realize, or are willing to admit. Analysts, project managers, copywriters, customer support reps, dispatchers, and so many more could all be replaced today. In cases where they can't be fully replaced overnight, you could certainly eliminate 90% of the workforce and leave the remaining 10% solely in charge of monitoring AI workflows. 

For many of us, the dream we were sold was fairly simple. We go to school, study hard, get into a good college, pick a field, and find a job in that field once we graduate. We were told that if we acquired skills and a degree, we would eventually find something stable. We could get a mortgage, a retirement account, and hopefully, if we were really lucky, a life that felt meaningful. This isn’t some cultural mythology I’ve created to serve my interest, this is what led policy in this country for decades. If you weren't born into a family of immense wealth and power, you had to work hard to earn yours. If you were born at the bottom of the socioeconomic ladder, the dream was always fairly similar. Get a degree, then get an entry-level job that lets you progress. This has been a blueprint for middle America since its existence. The middle class was never an accident, it was constructed. For most people reading this, you’ve probably built your life on the assumption that the foundation for this dream is still standing. You might have already progressed through this basic structure I’ve laid out, but I’m not sure it will be possible to do so going forward. 

There have been many revolutions in both technological and economic senses. Tractors put farmhands out of work, which led them to factories. Assembly lines gutted skilled craftsmen in every industry. Computers killed switchboard and telephone operators, typists, travel agents, hell, even grocery baggers and cashiers are becoming a rarity with self-checkout getting more popular. Losing jobs to technological innovation is nothing new, but this time, what if the new technology, AI and LLMs, stands to replace or eliminate more jobs than it could ever hope to create? What if AI doesn't just change what work looks like, but how much of it there even is? 

One thing to get out of the way early is that this is not a worker-centered movement. The top-level executives and billionaires funding this revolution are not sitting in their boardrooms or on their yachts worrying about how this will affect the paralegal in Cleveland or the train dispatcher in Fresno. They are not concerned about the growing displacement of the middle class, and if they are, it's not displacement that worries them. In 2025 alone, Apple, Microsoft, Alphabet, Nvidia, and other big tech companies collectively poured roughly $400 billion into AI infrastructure and development. They are not investing in your future, as was imagined at the beginning of the AI boom, they are investing in a future where you cost less, or more possibly, where you cost nothing at all.

Years ago, when people thought about AI automation, most assumed it would replace truck drivers and warehouse pickers. They assumed blue-collar, physical jobs that seem repetitive and “low-skill” were who AI was targeting, but as it turns out, the data tells a very different story. As it turned out, computer programmers, financial analysts, marketers, junior lawyers, and staff writers are first up to the AI chopping block. These are jobs AI can already do, and even better than you can. MIT researchers calculated that current AI systems could automate tasks performed by 20 million American workers. Much to the surprise of many of us, AI came after the cubicle, far before the warehouses and truck drivers. Make no mistake, though, warehouses and truck drivers are next. 

When mentioning some specific jobs that are getting replaced, or already have been, did you notice what kind of jobs those are? These are not “higher rung” jobs like top executives and board members, though those could be replaced too, it's overwhelmingly entry-level jobs. The kind of jobs 23-year-olds find themselves in when they’re trying to learn how industries really work. The kind of jobs where people network, gain relevant job skills, and learn what the corporate ladder looks like. Companies are already stopping entry-level hiring in fields with significant AI exposure. If junior analysts, staff writers, associate attorneys, and entry-level coders are all ceasing to exist, the bottom rungs of the corporate ladder are disintegrating, and the ladder has become purely decorative. If the ladder continues to get pulled up, who will the senior people even be in 10 years? The middle class doesn’t just need a job that covers their rent, they need a pathway for progression. If this trend continues, AI will be quietly dismantling the path that middle America has staked their lives and education on, while the end goal remains unchanged. 

AI is not like other industry disruptions, and we see that time and time again. In this revolution, companies benefit from a dramatic increase in productivity as they begin to incorporate AI into more jobs, but wages don't follow the same pace. According to the EPI, since 1979, worker productivity has grown by 64.6%, while wages have grown by only 17.3%, accounting for inflation. The extra profit being squeezed out of each worker, as always, goes to capital owners. What are those capital owners to do but reinvest into even more AI development? AI then has the power and capital behind it to make even more jobs redundant, which creates even more productivity for the companies, which creates even more profit for the capital owners, who invest it back into AI, which eliminates even more jobs, and on and on and on. This loop is not self-correcting, it’s compounding. The people at the top who started this loop, and who keep it spinning, have every financial incentive to keep it going as long as they can. The longer they ride this, the better their lives get, and the worse ours do. I’m not trying to push some grand conspiracy here, this is just math with serious consequences that no one in positions of power or wealth will ever have to personally feel.

The cruelest part of this whole scheme is who is absorbing these costs. It's not the firms announcing record margins, or the investors watching their AI portfolios compound, it lands on the people who did exactly the “right thing”, got their degree, took on the debt to pay for it, and are now watching their credentials lose their value faster than they can pay it off. For at least two generations, a college degree was the golden ticket. It said to employers, this person can be trained and comes pre-equipped with the skills they need to excel at their job. Employers paid better because they had this signal, but now, the tasks those people intended to do are all becoming automated. Now, overqualification has become the norm as people with marketing degrees fight tooth and nail over a $38,000 job. Student debt hasn’t changed, but the return on that debt evaporates year after year. The middle-class bet has always been on education, but that bet grows increasingly shaky every year, and the American dream is quickly becoming something you can’t quite afford to believe in anymore. 

For the people who are still holding onto their jobs and paying down their degrees, they’re beginning to feel it, too, when contract negotiations come back up again. The ambient knowledge that you could be replaced whenever your bosses feel like it has an immense impact, we’re only beginning to understand. Do you push for that next raise or promotion when you're not even sure you’re irreplaceable anymore? Would you leave a bad job earlier when the risk of changing industries or companies feels higher than it ever has? Of course not, you work even harder for the same amount of money as before because at least for now, you still have that money. Research shows that younger workers in AI-exposed fields are reporting lower job engagement and a weaker sense of purpose. The middle class isn't only some economic category you can fit neatly into a box, it's a psychological, cultural path that runs on confidence. That confidence is quickly corroding by the minute, and you don't need to be fired first for that to happen. 

The part of this that should terrify most of us is that consumer economics runs on the principle of volume, and that volume comes from the middle class having a disposable income. If you don't like one restaurant, that's fine, someone else does, and you can go somewhere else. Both businesses in this situation get to maintain some customers and profit. If this AI boom concentrates wealth at the very top, as wealth usually does, and then guts the purchasing power of everyday Americans, you get an abundant dystopia. You get an abundance that no one can afford. We now have “lights-out” factories and warehouses, that is, factories run without any human involvement at all. I'm sure if you don't pay any labor costs at all, your quarterly reports in your boardrooms are exciting events, that is, until demand collapses. Even Henry Ford understood this, even though he wasn’t being incentivized to by other capital owners. He had a revolutionary idea to pay his workers enough to buy the cars they made, and that logic hasn't changed, but the people ignoring it have gotten very good at externalizing the consequences. If all cars cost $1, what's that really worth if you're not making any money at all, and the prospect of making $1 has gotten even more far-fetched? Elon Musk will not buy 50,000 cars to keep car companies operating, and Bezos will not buy 100,000 steaks a day to keep restaurants open. These capital owners cannot sustain our economy on their own, only we can.

Here’s what the billionaires and hedge funds funding this revolution really don't want you to know. These outcomes, this doomsday scenario I'm painting, hasn’t been written in stone yet. The loop continues to run because they are propping it up and seeing how fast it can spin, regardless of the consequences. There still are regular people, whether that be economists, labor organizers, or a very select handful of legislators who haven’t fully given up yet, and have been trying to sketch out an off switch for years. Things like a universal basic income serve as a floor for financial stability and help mitigate some of the most severe financial insecurities this new AI boom could deepen. Bernie Sanders and others have been vocal about introducing a 32-hour workweek to find a new way to distribute work rather than just eliminate workers. Germany has led the charge with mandatory worker seats on corporate boards for companies above a certain size, an idea that would weaken the corporate elite and close-door board meetings. Legislators have introduced the idea of an AI or robot tax that would redirect productivity gains back into public systems, or the people AI seeks to replace, and profit sharing that is inherently tied to automation adoption. I'm not trying to sell you on all or really any of these ideas, but knowing they exist matters. The alternative framing, the one the executives have already decided on, is that nothing can be done about this. They believe this was an inevitable moment for human history, and everyone better get on board or go live in the streets. That’s not true, and we should all say so as clearly as we can. 

The people in power and the capital owners are counting on you not knowing any of that. They’re counting on the argument feeling too fixed, and too far above your pay grade. It isn’t.     I keep coming back to the most mundane job I ever had, in the plating room. If it can happen there, in that unremarkable corner of an unremarkable industry, it can happen anywhere, to anyone. It's no longer a question of whether it will come for your job anymore, it's just a matter of when. Are you ready to ditch the industry you’re in and have spent 30 years and hundreds of thousands of dollars equipping yourself for? If you were lucky enough to retire before any of these decisions affected you, I envy you, but your children and grandchildren won't be nearly as lucky. I’m not trying to give the impression that AI will become some evil Skynet-like tool, and that we should all live in fear of what it will do to us, but AI is a tool, and like all tools, it serves someone’s interests. Right now, no one is using these tools in service of you or me, but rather they are being used by the corporate elite to make our very work lives redundant. The question has always been deceptively simple, and it's not whether this changes everything; we already know it will. The question is, who is that change for? This answer has not been set in stone yet, so it still matters what we demand. Let's demand much more than survival. Let’s demand that this tool be used to serve us, the middle class, above all else, not stock prices or billionaires.