For decades, Hollywood warned us about killer robots marching through the streets, reducing humanity to piles of ash. But that scenario represents only one side of the rather scary story.

The alternative is that humans remain firmly in control of advanced technology. This is a more reassuring prospect for many, I imagine, but not for me. That’s because the age of AI risks creating a form of high-tech feudalism, in which people rent access to essential tools but never own them.

If that sounds alarmist, consider social media’s dry run from 2012 to 2023. Centers for Disease Control and Prevention data revealed a sharp spike in teen depression during that decade, remarkably coinciding with the rise of algorithms that seemed to know everything about everyone. In 2021, internal documents leaked from Meta proved that company executives knew their design choices were harming the mental health of young users. Engineers kept the engagement loops running anyway.

The reason was simple: User attention translates into ad revenue, and more attention means more money. An entire generation grew up trading its sanity for likes, clicks and shares.

Things were bad, but they’re likely to get worse. Social media merely hijacked our attention. AI is coming for the rest of the operating system. Generative models construct synthetic realities and personalized companions that mirror human affection. Today, a teen can spend their entire waking hours talking to a custom conversational bot that remembers every secret and gradually replaces real human interaction. AI monopolizes human cognition and influences how people think and act. Users become entirely dependent on proprietary software for even the most basic tasks, including sending emails and texting a loved one.

As the machines get smarter, we get dumber. And the people building the machines profit handsomely from the madness.

OpenAI is now worth $852 billion. Microsoft committed more than $13 billion to secure commercial access to its underlying models. Anthropic, Meta, Alphabet, and Amazon spend tens of billions annually on computing infrastructure to maintain market domination. These entities operate to maximize shareholder value above all other priorities. Staff members who voice legitimate concerns are often ignored or shown the door. Wall Street demands exponential growth. Executives deliver that growth by locking users into addictive and exploitative ecosystems.

In 1998, major tobacco companies signed the Master Settlement Agreement after decades of knowingly selling addictive products to teenagers. In 2008, global investment banks crashed the mortgage market and collected taxpayer bailouts while millions lost their homes. History shows a consistent pattern of corporate actors pursuing immediate profits while shifting long-term societal costs onto the public. Why would anyone of sound mind expect that to change?

Tech executives ask the public to trust their internal ethics committees, even as many figures inside the industry voice deep concerns about where this technology could lead. These are the people building the systems themselves. They understand that a small number of companies could eventually control technologies capable of doing a vast share of the work humans do today. If that sounds like an exaggeration, consider the pace of progress. AI is advancing rapidly. It’s not unreasonable to imagine a future, perhaps within the next decade, when these systems outperform humans across virtually every task.

To compound matters, the days of the plucky little startup are coming to an end. The sheer cost of building powerful AI systems creates an almost insurmountable barrier to entry. Running an advanced model requires specialized data centers that consume hundreds of megawatts of power and billions of dollars in hardware. A normal American operating out of a garage doesn’t stand a chance. The centralized setup guarantees that a select group of tech executives will maintain absolute control over the digital infrastructure of modern civilization. Every search and every automated workflow will pass through servers owned by four or five mega-corporations.

That dominance over information creates an unprecedented source of influence on society. The word unprecedented matters. Comparisons to the Industrial Revolution or the internet are nonsensical. Neither put this much power over human knowledge and decision-making in the hands of so few private companies.

The dystopian scenario requires no Terminators, only compliant consumers paying monthly bills to let their friend Claude tell them what to say and how to say it.