Generative AI Won’t Boost Productivity Overnight, Fed Warns Despite Long-Term Potential

Generative AI is not merely a passing technological trend but a transformative force with the potential to significantly boost human productivity over the long term, according to a new paper from the Federal Reserve. While the benefits are expected to be substantial, the path to realizing them will be slow, complex, and filled with risks. The Fed’s analysis places generative AI in the same category as historically transformative technologies such as electricity, the computer, and the microscope—what economists call general-purpose technologies. These innovations don’t just offer immediate efficiency gains; they spark continuous waves of follow-on innovation and reshape entire economies over decades. The paper identifies two key types of impactful technologies. The first is general-purpose technology, which drives sustained productivity growth even after widespread adoption. Examples include the electric dynamo and the digital computer. Generative AI shows strong signs of fitting this mold. From specialized models like OpenAI’s LegalGPT to AI-powered tools such as Microsoft’s Copilot, AI is already being embedded into workflows across industries. The rapid pace of advancement—evident in models like Deepseek’s R1 and the rise of agentic AI—suggests that this is just the beginning of a broader innovation cycle, particularly led by digital-native firms. The second category is “inventions of methods of invention,” such as the printing press and the microscope. These tools don’t just improve existing processes—they unlock entirely new frontiers of discovery. The Fed highlights AI’s growing role in scientific research, including drug development and simulations of complex systems like the universe. Since 2023, there’s been a sharp increase in companies referencing AI in research and development and in earnings calls, signaling early integration into corporate innovation strategies. Despite this optimism, the report underscores a critical reality: the productivity gains from AI will not come quickly. The main obstacle isn’t the technology itself, but adoption. While some sectors, especially finance and tech, are actively integrating AI, most businesses—particularly small ones—have yet to incorporate it into daily operations. Even among researchers, uptake is still limited. The Fed notes that full productivity benefits will require complementary advancements in user interfaces, robotics, AI agents, and infrastructure. These supporting systems take time to develop and deploy. The timeline for a measurable productivity boom remains uncertain, though some analysts, including those at Goldman Sachs, project meaningful impacts on U.S. GDP and labor productivity beginning around 2027, peaking in the 2030s. Another risk lies in overinvestment. As demand for AI grows, so does the need for data centers and energy. The Fed warns that building infrastructure too quickly—without clear demand—could lead to wasted capital and economic strain, much like the railroad overexpansion of the 19th century. In conclusion, the Federal Reserve sees generative AI as a long-term game-changer for productivity, comparable to past revolutions in technology. But its full potential depends not on the AI itself, but on how quickly and effectively organizations, workers, and governments adapt to it. The transformation is coming—but it will take time.