The Merchants of the AI Apocalypse: Silicon Valley CEOs Are Selling Americans Their Own Fictional Destruction for $20 a Month

Your Extinction Is $20 a Month: Silicon Valley’s strangest business model may also be its most profitable, spend years warning Americans that artificial intelligence could take their jobs, destabilize society and potentially threaten humanity itself, then charge them $20 a month to use it. The technology may become indispensable infrastructure, but that doesn’t mean the public will continue trusting the executives who sold it through fear.

The New Robber Barons: When Utility Outruns Trust

The closest historical parallel to the strange relationship Americans are developing with artificial intelligence may not be the internet, the smartphone or even the broader Industrial Revolution. It may be the railroad boom of the 19th century, when a revolutionary technology became indispensable to American economic life while many of the men controlling it became symbols of concentrated corporate power.

In the decades after the Civil War, railroads transformed the United States, connecting farms to cities, manufacturers to national markets and previously isolated communities to the broader economy. The scale of the expansion was extraordinary: federal historical statistics show the United States had roughly 30,600 miles of railroad in 1860 and more than 163,000 miles by 1890. The technology was not merely popular; it was becoming national infrastructure. Yet growing dependence on railroads did not translate into public affection for railroad corporations or the powerful financiers associated with them.

That distinction matters enormously in the AI era. The political backlash against railroad power became particularly intense among farmers and small businesses that depended on rail transportation but believed discriminatory rates, rebates and concentrated control left them vulnerable to railroad companies. The resulting Granger movement pushed states toward railroad regulation, while Congress eventually passed the Interstate Commerce Act of 1887, establishing the Interstate Commerce Commission and making railroads the first American industry subjected to federal economic regulation.

The country did not respond to railroad abuses by tearing up the tracks. It regulated the companies controlling them. That is the historical warning Silicon Valley should be studying.

People Can Love the Product and Distrust the People Selling It

The railroad comparison is useful precisely because it separates technological utility from institutional legitimacy. A farmer could depend on a railroad to reach distant markets while simultaneously resenting the railroad company setting the freight rate. A merchant could recognize that trains represented extraordinary technological progress without believing railroad executives should have unlimited power over commerce. AI appears capable of producing the same contradiction.

Millions of people can find generative AI extraordinarily useful for writing, coding, research, translation, customer service, data analysis and dozens of other tasks without automatically trusting the executives building the systems. Adoption is not an opinion poll. Usage does not necessarily constitute approval, and dependence certainly does not confer legitimacy. That distinction becomes even more important as AI migrates from stand-alone chatbots into search engines, office software, smartphones, operating systems and enterprise workflows. The more deeply AI becomes embedded in ordinary digital infrastructure, the less meaningful raw usage numbers become as measurements of public enthusiasm.

Recent polling underscores that disconnect. Pew Research Center reported in June 2026 that roughly six in ten American adults lacked confidence in U.S. companies to develop and use AI responsibly, while 67% expressed little or no confidence in the federal government’s ability to regulate AI effectively. Another Pew survey conducted in June 2026 found 52% of Americans were more concerned than excited about the growing use of AI in daily life, compared with only 9% who were more excited than concerned. A technology can become increasingly useful long before the institutions controlling it become trusted.

The AI Industry Has an Additional Problem the Railroad Barons Never Had

The historical comparison breaks down in one particularly revealing place: the sales pitch. Jay Gould did not promote railroads by repeatedly suggesting locomotives might eventually become intelligent enough to exterminate humanity.

Modern AI leaders have operated inside a far stranger communications environment. Some of the industry’s most prominent executives and researchers have publicly warned about catastrophic risks from sufficiently advanced artificial intelligence while their companies simultaneously race to develop more capable models, attract investment, acquire computing infrastructure and embed AI throughout the global economy.

That tension reached an extraordinary point in May 2023, when the Center for AI Safety published a one sentence statement declaring that mitigating the risk of extinction from AI should be treated as a global priority alongside pandemics and nuclear war. Signatories included OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google DeepMind CEO Demis Hassabis, Geoffrey Hinton, Yoshua Bengio and numerous other prominent researchers and technology executives.

Whatever one believes about the underlying risk, the statement crystallized an almost unprecedented corporate communications problem: some of the people building the technology were simultaneously telling society that increasingly powerful versions of that technology could present catastrophic dangers. Imagine a railroad executive appearing before Congress in 1885 and saying his locomotives might eventually become powerful enough to destroy civilization, before returning home to announce plans to lay tracks through every major American city. That is approximately the psychological contradiction the AI industry has asked the public to process.

The Reverse Marketing Problem

There is a legitimate debate over whether existential risk warnings represent responsible disclosure, sincere scientific concern, strategic positioning, regulatory maneuvering or some combination of those factors. Those possibilities should not be casually collapsed into a single explanation. But the effect of the messaging on public trust can still be examined.

Traditional technology marketing generally emphasizes agency: this machine will make you faster, smarter, more creative, more productive or more connected. The personal computer industry sold empowerment. The smartphone industry sold convenience and connection. AI has frequently arrived accompanied by a considerably darker conversation: the technology could eliminate jobs, destabilize professions, supercharge misinformation, become increasingly difficult to control and according to some of its own developers, potentially present catastrophic risks.

Then the same industry asks everyone to integrate it into their lives.

That is not conventional marketing. It is closer to selling electricity by repeatedly reminding customers that the power plant might someday burn down the city.

The rhetoric unquestionably established artificial intelligence as one of the most consequential technologies in the minds of investors, policymakers and the public. But it also risks creating a population that simultaneously depends upon AI and distrusts the institutions controlling it. That is where the railroad comparison becomes uncomfortable.

Infrastructure Eventually Becomes Boring

The most disruptive technologies frequently become psychologically smaller as they become technically larger. Electricity was once a technological wonder; today people flip a switch. Railroads transformed geography before becoming transportation infrastructure. The internet was once discussed as “cyberspace,” almost as though humanity were entering another dimension; today someone orders dinner through it without thinking about TCP/IP.

Artificial intelligence may undergo the same demystification. Large language models are extraordinarily sophisticated computational systems, but they remain engineered systems operating on physical hardware, consuming electricity, executing mathematical operations and producing outputs in response to inputs. As AI becomes familiar, the mythology surrounding it may shrink even as its practical importance expands.

That could create a difficult moment for the industry’s leadership. If ordinary users eventually come to regard AI less as an approaching synthetic deity and more as another layer of computing infrastructure, they may revisit years of apocalyptic rhetoric with considerably less patience.

If this was always going to become ordinary software infrastructure, why were the people selling it constantly telling us to contemplate the end of civilization?

There may be legitimate answers to that question. Advanced future systems could present risks dramatically different from today’s models, and researchers are entitled to warn governments and the public about dangers they genuinely believe deserve attention. But public trust depends not merely on whether a warning can be defended. It depends on whether institutions appear consistent, transparent and worthy of the enormous authority they are asking society to grant them.

The Lesson of the Railroad Age

The strongest historical lesson is not that today’s AI CEOs are identical to Jay Gould, Cornelius Vanderbilt or other Gilded Age industrialists. They are not, and forcing the analogy that far would turn useful history into caricature. The more important parallel is structural: a society can enthusiastically adopt a technology while becoming increasingly skeptical of the concentration of economic and institutional power surrounding it.

Americans ultimately kept the railroad network while imposing increasingly substantial political and regulatory constraints on railroad corporations. The infrastructure survived the backlash against the institutions controlling it. AI could follow a comparable trajectory.

The models may become permanent. The APIs may disappear invisibly into thousands of products. AI assistants may eventually become as unremarkable as search bars. Businesses may become increasingly dependent on automated systems. None of that guarantees that today’s executives, corporate structures or concentrations of power will retain public legitimacy. That is the mistake every generation of technological elites risks making: confusing dependence with admiration.

People can need your product and still distrust you. They can recognize that your technology changed the world and simultaneously decide that the people controlling it have accumulated too much power. The railroad barons discovered that distinction more than a century ago. Silicon Valley may be discovering it again.

Sources & Further Reading

U.S. Census Bureau — Historical Railroad Mileage Statistics

National Archives — Interstate Commerce Act of 1887

Center for AI Safety — Statement on AI Extinction Risk

Center for AI Safety — May 2023 Statement and Signatories

Pew Research Center — Americans and AI 2026: Chatbots, Smart Devices and Views on Impact

Pew Research Center — Young Americans Are Increasingly Wary of AI