Over 200 economists and AI researchers, including 16 Nobel laureates, have signed a statement urging immediate action on artificial intelligence, warning that its evolution could transform the global economy far more swiftly than the Industrial Revolution. Among the signatories, Marc Esposito, a technology policy professor at Northeastern University and a senior economist at micro1, along with Auralie Jean, CEO of the AI-driven women’s health startup Infra, emphasize the imperative to guide AI towards complementing humans rather than replacing them.
The statement highlights a critical distinction: AI should enhance human capabilities rather than merely replicate them. This differentiation is not just theoretical; it carries significant economic implications. AI can augment human potential by expanding access to valuable resources and reducing operational costs, thereby enabling people to achieve what was once beyond their reach. Conversely, if AI is limited to automating specific tasks, it merely redistributes existing value rather than expanding economic opportunities.
A prevalent but flawed economic notion suggests there is a finite amount of work available. This “zero-sum” mindset assumes that every automated task displaces a human worker, posing the only question of how rapidly technology replaces jobs. Historical precedents, however, illustrate otherwise. Innovations such as the loom, tractor, and personal computer were expected to induce widespread unemployment. Instead, they lowered production costs and created more jobs than they eliminated.
As Erik Brynjolfsson, a key figure behind the “We Must Act Now” campaign, has termed it, the “Turing Trap” results from designing AI to mimic humans rather than enhance their roles. This tendency consolidates power among the owners of AI systems and limits opportunities for ordinary workers to create value. While imitation merely divides existing economic “pie,” enhancement has the potential to enlarge it.
The mechanisms by which technology drives economic growth are well established but frequently overlooked. When automation reduces production costs, demand generally increases, leading to a higher demand for human labor in roles that machines cannot fulfill. A case in point is the introduction of ATMs, which initially reduced the number of tellers in each bank branch from about 20 to 13. However, the lower operating costs allowed banks to open more branches, resulting in a sustained growth of teller jobs as the role evolved from counting money to providing customer service.
In a 2019 research paper, economists Daron Acemoglu and Pascual Restrepo provided a framework for understanding these dynamics. They explained that while automation transfers certain tasks from humans to machines, the resulting productivity gains are often accompanied by the creation of entirely new roles that require human skills. Many professions that have driven job growth in recent decades barely existed a generation ago.
However, there are no guarantees of positive outcomes. Acemoglu and Restrepo also described “jobless automation,” where technologies replace workers without generating sufficient productivity gains or creating enough new demand to foster new opportunities, thereby diminishing labor’s share of the economy.
AI has the potential to reshape the economy on an even larger scale than past waves of automation. Instead of simply automating individual tasks, AI seeks to lower the cost of insight itself. As the cost of such general-purpose inputs declines significantly, markets not only consume more of the same services but also innovate new ones. Cutting the costs of legal, medical, educational, and scientific expertise—fields previously constrained by a shortage of skilled professionals—enables broader access without necessarily replacing specialists.
This embodies true integration; the tool complements humans by enhancing judgment, taste, accountability, relationships, and other uniquely human capabilities that AI cannot replicate. Increased availability of raw cognitive abilities elevates, rather than diminishes, the value of distinctive human contributions.
Of course, if these outcomes were guaranteed, the statement we signed would be unnecessary. The challenge of AI either augmenting or substituting for human labor hinges on the choices we make today. While it is straightforward to calculate the cost of replacing a worker and justify that replacement, creating tools that empower people to achieve previously unimaginable feats is far more complex and time-consuming, yielding broader benefits.
To unleash the full economic potential of AI, it must be directed toward expansion rather than replacement. This shift requires replacing incentives that favor capital at the expense of labor with those that reward and invest in enhancing human capabilities. Success should not be measured solely by the number of jobs AI replaces, but by the scale of new demand and opportunities it generates. An AI system that focuses merely on automating existing tasks fails this test, even if it leads to short-term profit increases.
As emphasized earlier, the question is not whether AI will grow powerful—it already is and will continue to gain strength—but whether we will harness this power to narrow the scope of human work or expand it into realms we have yet to envision. Instead of merely automating tasks humans already perform well, we should empower them to accomplish the previously unthinkable and ensure that the resulting prosperity is widely shared.
