For the past two years, a steady drumbeat of headlines has warned us to prepare for a “job apocalypse.” From Silicon Valley boardrooms to the halls of Congress, the narrative was settled: generative artificial intelligence would hollow out the middle class, erase entry-level careers, and leave a trail of mass displacement in its wake. We were told that the cognitive revolution would succeed where the industrial revolution failed — by finally making the human worker obsolete.
But as the dust settles on the first wave of adoption, we finally have something the doomsayers rarely contend with: actual data, not just dark projections.
New research from Vanguard reveals that the “Great Displacement” is actually a “Great Upgrade.” Contrary to the prevailing panic, both jobs and wages are growing significantly faster in professions most exposed to AI than in those with little to no exposure. Between 2023 and 2025, real wages in highly AI-exposed roles climbed 3.8%, while less exposed roles stagnated at a mere 0.7%. Far from triggering a collapse, AI is doing what transformative tools have done for centuries: supercharging worker productivity and, by extension, worker value.
Rather than triggering a jobs collapse, AI is doing what transformative technologies have long done. It is dramatically improving workforce productivity. This increased efficiency is resulting in new economic opportunities.
This data-driven reality stands in sharp contrast to the fear-based rhetoric that has dominated headlines. When Anthropic’s CEO Dario Amodei predicted AI will “wipe out” half of entry-level jobs, we should look closely at who benefits from that fear. Such apocalyptic language may generate clicks, but it also serves a convenient corporate purpose: inviting heavy-handed regulation that acts as a moat, protecting deep-pocketed incumbents from the very startups that would use AI to create the next generation of jobs. It is regulatory capture wrapped in a moral panic.
Many AI policy observers recognize this pattern for what it is: regulatory capture wrapped in moral panic. News aggregators and pundits, meanwhile, eagerly amplify phrases such as “jobs apocalypse,” “end of work as we know it,” and “human obsolescence.” Outrage sells. Evidence struggles to compete.
History shows this cycle is nothing new.
In 1961, a majority of the public opposed the race to the Moon, dismissing it as a costly “moondoggle.” During the 1930s, Albert Einstein blamed automation for the Great Depression, warning of “great distress” caused by machines. Earlier still, critics predicted the bicycle would collapse marriage rates, destroy book sales, and depress furniture purchases.
Seen in this context, today’s AI panic looks less prophetic and far more familiar.
What is different now is that we no longer have to rely on historical analogy alone. We have contemporary evidence. When researchers do examine the data closely, the results are strikingly non-apocalyptic. The Vanguard report actually builds off of other recently published analyses that came to the same conclusion.
Recently, Yale Budget Lab authors found no discernible economy-wide disruption since ChatGPT’s release 33 months ago, undercutting fears that AI is eroding demand for cognitive labor.
Not to be outdone, MIT’s Lawrence Schmidt demonstrated that from 2014 to 2023, AI adoption was associated with workers shifting into higher-value, more creative, and analytical roles rather than being pushed out of the workforce.
There is, in fact, reason to be hopeful that more tech transitions will create new opportunities. As Box CEO Aaron Levie recently explained, “I don’t think the world has enough engineers building software for the next set of problems that are going to exist.” New challenges create new vocational opportunities.
None of this is to suggest that AI poses no challenges. The technology is advancing rapidly, and its long-term effects will depend on policy choices, business practices, and how broadly productivity gains are shared. But governing from a place of fear, or mistaking speculative worst-case scenarios for present-day reality, is a recipe for stagnation rather than security.
The lesson from the growing body of empirical research is clear. Technological change is not destiny, and panic is not policy. AI will reshape work, just as past innovations have. The question is whether we respond with superstition and pessimism or with confidence grounded in evidence.
Data, not fear, should guide the national conversation. The United States is once again on the cusp of a major productivity transformation. If we meet the moment with clear eyes, AI can be not a sign of doom, but the next chapter of American opportunity.
Nathan Leamer is the Executive Director of Build American AI, a 501(c)4 nonprofit organization working to advance U.S. policy leadership in artificial intelligence.
