Today, 200 economists, including sixteen Nobel laureates, signed a joint statement warning that AI could drive an economic shift larger than the Industrial Revolution, but in a fraction of the time. The statement is four sentences long. It offers no specific policy proposals and no timelines. What it does offer is an admission: the people whose job it is to understand economic change are telling you, plainly, that they currently cannot.
The statement was coordinated by Stanford's Digital Economy Lab. The signatories include former Google CEO Eric Schmidt, LinkedIn co-founder Reid Hoffman, Nobel laureates Joseph Stiglitz, Paul Krugman, and Daron Acemoglu, alongside people from Google, OpenAI, and Anthropic. Getting economists who routinely disagree to sign the same document is genuinely rare. Acemoglu, for instance, has spent years arguing that AI's productivity benefits are overstated, while others on the list are considerably more optimistic. The shared concern is not the direction of impact, but the speed.
The core problem the statement points to is not that AI will destroy jobs. It is that the economic tools used to track and respond to major shifts are far too slow for what may be coming. Tom Cunningham, one of the statement's organizers, put it plainly: "We are driving in fog." That fog is institutional, not just theoretical.
The data underneath the statement tells a split story. A Federal Reserve study found that US programmer job growth was cut roughly in half after ChatGPT launched in late 2022. Over three years, researchers estimated a shortfall of around 500,000 roles that would have existed under pre-AI growth trends. The gap only became visible in mid-2024, about eighteen months after the tool launched, suggesting companies tested AI tools quietly before adjusting headcount.
At the same time, broader labor market data shows no mass disruption. The Yale Budget Lab found no significant changes in occupational mix or unemployment levels in high-AI-exposure jobs. The St. Louis Federal Reserve found no clear evidence of either job gains or losses at the industry level from AI adoption. Separate IMF research found AI adoption is still concentrated among a minority of workers, meaning the technology has not spread widely enough to produce broad disruption yet.
The pattern that does emerge from the data is more targeted than headlines suggest. Entry-level roles in white-collar work are shrinking first. Big Tech new graduate hiring fell nearly 50% from pre-pandemic levels. Top law firms slowed first-year associate intake. The Big Four accounting firms restructured entry-level audit positions. Anthropic CEO Dario Amodei warned a year ago that AI could eliminate half of entry-level white-collar jobs within five years, naming finance, law, consulting, and technology specifically. He has since softened that view somewhat, now describing AI more as a productivity multiplier than a job eliminator. But the entry-level data has continued moving in the direction he described.
This is the detail that matters most for business operators: the disruption is not arriving evenly. It starts at the bottom of the career ladder, in roles defined by routine, digital, well-structured tasks. Junior analysts, first-year associates, entry-level coders, and new-hire consultants are facing reduced demand, not because their companies are shrinking, but because more senior people with AI tools can now cover what those roles used to do. That has two consequences. It reduces hiring costs in the short term. It also removes the training pipeline that produces senior people a decade from now.
The statement asks for research, institutions, and policy guardrails. It does not specify what those should look like. The honest read is that the people signing it do not yet know what the right guardrails are. What they do know is that waiting for certainty is itself a choice, and probably not the right one. For any organisation with significant white-collar headcount, that is the number worth sitting with.