In May 2025, Dario Amodei, the CEO of Anthropic, predicted that artificial intelligence could eliminate roughly half of all entry-level white-collar jobs within five years, with unemployment potentially spiking to twenty percent. A few weeks later, Shopify's CEO told his employees that no team would be allowed to request additional headcount until they could demonstrate the work couldn't be done by AI. Standard Chartered announced it would cut almost eight thousand positions as part of an AI-driven strategy. Across the industry, the message was consistent, deliberate, and unambiguous: this technology will replace you.

Then the industry acted surprised that people believed it.

A Wall Street Journal analysis published this month reports that artificial intelligence is now less popular with Americans than ICE, less popular than politicians, a threshold most observers would have considered physically impossible to reach. Communities have blocked or delayed at least forty-eight data center projects worth $156 billion. Voters in Festus, Missouri ousted four city council members a week after they approved one. Someone threw a Molotov cocktail at Sam Altman's house. Roughly 360,000 Americans have joined Facebook groups opposing AI facilities, which is approximately 360,000 more people than have joined Facebook groups in favor of anything since 2019.

The AI industry has a backlash problem. But the backlash didn't come from nowhere. The industry built the narrative that produced it.

The fear narrative served three distinct audiences, and each one heard a different frequency.

For investors, existential capability claims inflated the perceived total addressable market. If AI can replace half the white-collar workforce, the revenue opportunity is measured in trillions. The capital followed accordingly: AI captured eighty percent of global venture capital in the first quarter of 2026. The more powerful you claim your technology is, the more capital you attract. Fear of displacement and awe at capability are the same claim, arriving at different ears.

For corporate boards, AI provided a clean justification for conventional cost restructuring. In his annual report last year, Standard Chartered's CEO Bill Winters praised the "extraordinary commitment and ingenuity" of his 82,000-strong workforce, predicting that their "collective spirit and drive" would define the bank's next chapter. This week, announcing eight thousand job cuts, he explained the bank would be replacing "lower-value human capital" with AI. It is, as they say, a journey. Winters has since rowed back the comments, though one suspects the 8,000 people clearing their desks found the clarification less reassuring than the annual report. Shopify didn't freeze headcount because it had overhired. It told employees to prove AI couldn't do their jobs first. You could think of it as AI-washing: the job losses were coming anyway, but dressing them in AI language converts a defensive cost story into a forward-looking strategic one. It's the difference between telling your board "we're cutting costs" and telling them "we're investing in the future." Both mean the same thing for the lower-value human capital.

For regulators, the existential framing shaped the policy conversation on favorable terms. If your technology is powerful enough to eliminate half of all entry-level jobs, the appropriate government response is to help manage the transition, not to question whether the premise is real. Every prediction of mass displacement doubled as an argument against precautionary regulation. You don't regulate a weather system. You build shelter.

The problem was audience containment.

The narrative was designed for boardrooms, earnings calls, and Davos panels. It reached living rooms, school gates, and city council meetings. And the people who received it (unlike investors, who heard opportunity, or board members, who heard strategic cover) heard exactly what the words said. The TUC's Adam Cantwell-Corn summarized the public's understanding neatly: the current narrative is that AI will take your jobs and may end up killing us. Hard to imagine why that's not polling well.

Research published this month by Bobby Duffy at King's College London quantifies the gap. Sixty-nine percent of businesses are optimistic about AI's impact on employment. Twenty-eight percent of the general population share that optimism. The majority of all groups surveyed, including employers, employees, young people, and students, predicted that AI's economic benefits would flow mainly to wealthy investors and large companies rather than to workers or society at large. A Gallup poll found that seven in ten Americans oppose data center construction "to support artificial intelligence." Remove the words "artificial intelligence" and the result would likely reverse. The technology isn't what people object to. It's the story they've been told about what the technology is for.

You cannot spend three years telling the market your technology is the most disruptive force since electricity and then pivot to reassuring communities that it's perfectly safe and will create more jobs than it destroys. The market heard the first message because it was profitable to send it. The public heard it because the public has ears.

There is a specific pathology at work here, and it goes beyond marketing miscalculation.

The AI industry has adopted what you might call the Oppenheimer posture: positioning its own creation as so powerful it might destroy civilization, while simultaneously arguing that it should be the one trusted to manage the consequences. "Now I am become Death, the destroyer of worlds" was a statement of horror. In Silicon Valley, it has become a pitch deck.

The incentive structure rewards this posture at every level. Founders who claim civilization-scale impact attract more capital than founders who claim incremental productivity improvements. CEOs who frame layoffs as AI-driven transformation get better market reactions than CEOs who admit to conventional restructuring. The entire industry has a structural incentive to overstate its own danger, not because the technology isn't genuinely powerful, but because the narrative of power has become disconnected from the evidence of impact. Eighty-eight percent of companies now use AI in at least one function. Only around forty percent report positive bottom-line impact. Companies are announcing AI-driven layoffs before the technology has delivered the productivity gains that would justify them, which is a bit like selling your furniture because you've heard someone invented a chair.

Amodei himself appears to have noticed the problem. As of this year, he has pivoted from predicting that AI will eliminate half of entry-level white-collar jobs to arguing, via Jevons' Paradox, that AI will ultimately create more work than it destroys. The narrative adjusts when the political invoice arrives. Fear was useful when it attracted capital and shaped regulation. It becomes a liability when it blocks your data centers and turns your commencement speeches into protest events.

The industry's response to the backlash has been, so far, instructive. OpenAI's Chris Lehane blames "doomers" and negative media coverage for driving public fear. One executive at a Washington data center conference dismissed opponents as "cave people." President Trump suggested data centers "need some PR help," an understatement from a man not generally known for them. Lehane says the industry needs to be "more calibrated in making the case." He's right, though probably not in the way he intends. The case was extremely well calibrated: it attracted record capital, justified corporate restructuring, and shaped the regulatory conversation. The calibration failure wasn't in the message. It was in the assumption that a message designed to impress investors wouldn't terrify everyone else.

What makes this doubly wasteful is that there is a genuinely compelling answer to the question "why should we welcome this technology?", one the industry has barely tried to articulate, because it requires more imagination than the cost-reduction story and doesn't fit on an earnings-call slide.

The gut-check reaction to AI in most companies, the one that shows up first in boardroom conversations, the one that requires almost no original thinking, is to use it to eliminate jobs. In customer experience, the most obvious version is the contact center: replace human agents with AI, cut headcount, report the savings. The technology arrives, you point it at your most visible cost line, and you book the reduction. Whether customers end up with a better experience is, at best, an afterthought. Whether you've created any actual economic value is a question nobody seems to be asking, which is convenient, because the answer in most cases is no. You've moved a cost from one line to another and called it a strategy.

Now consider the alternative. AI applied not to eliminating the people who talk to customers, but to understanding the customers themselves: what they're likely to do, where the growth opportunities sit, which accounts are at risk before anyone has complained, which operational patterns predict expansion and which predict departure. This is AI as a growth engine rather than a cost engine. It uses data and intelligence to find revenue that would otherwise go undetected, to prevent losses that would otherwise go unnoticed until they showed up in the quarterly numbers. It demands a theory of how the business actually grows, not just a line item to cut. But it presents a fundamentally different vision of what the technology is for, one that creates economic value rather than merely redistributing it from employees to shareholders.

The fear narrative, that AI will take your job, is the natural consequence of an industry that has chosen to lead with cost reduction. If the primary use case you're selling is headcount elimination, you shouldn't be too surprised when the heads in question object. But if the story were about growth (finding customers, keeping customers, understanding what actually drives the economics of a business) the conversation changes. Not because the technology is different, but because who benefits is different.

The fear-mongering isn't just politically self-destructive. It represents a failure of imagination. The industry built a fear story because a fear story was easier to sell than a growth story. The growth story requires you to think. The fear story just requires you to count.

I'm Richard Owen, founder and CEO of OCX Cognition. We build predictive customer analytics for companies who'd prefer to know which customers are at risk before those customers have already decided to leave.