A new Brookings Institution study warns that artificial intelligence could initially boost wages by enhancing worker productivity, but may ultimately drive them down as automation replaces human roles in cognitive tasks. The research, led by Konrad Kording, a professor of neuroscience at the University of Pennsylvania’s Integrates Knowledge (PIK) program, and Ioana Marinescu, an associate professor at Penn’s School of Social Policy & Practice, presents a nuanced view of AI’s long-term economic impact. The study models the transition from human-led to machine-led intelligence and finds that automation in the intelligence sector follows a hump-shaped trajectory: wages rise sharply at first due to increased productivity, then plateau, and eventually decline as AI systems take over more tasks. This pattern reflects a temporary boom followed by a correction, as the benefits of automation increasingly flow to capital owners rather than workers. The researchers explain that while AI initially increases demand for skilled labor by amplifying human capabilities, over time, as machines handle more complex cognitive work, the need for human input in those areas diminishes. Workers are then pushed into slower-growing, physical labor sectors—such as construction, transportation, and caregiving—where productivity gains are limited and wages stagnate or fall. Despite continued growth in overall output, the study shows that wages can decline even as economic activity expands. This divergence highlights a growing imbalance: digital productivity surges, but labor’s share of that growth shrinks. The authors reject both extreme views of AI’s future—the utopian vision of endless abundance and the dystopian fear of mass unemployment. Instead, they introduce the concept of “intelligence saturation,” which suggests that AI’s gains are constrained by real-world limits: physical infrastructure, human oversight, and the need for tangible tools and equipment. To prevent wage declines and protect workers, the researchers advocate for a more measured rollout of automation. They recommend investing in physical capital—machines, tools, and infrastructure—to ensure that human labor remains valuable even as digital tasks are automated. They also propose taxing virtual substitutes for in-person services, such as AI-driven consultations or remote care, to discourage companies from replacing human workers with machines. This idea echoes earlier calls, including Senator Bernie Sanders’ proposal for a “robot tax,” aimed at curbing job displacement. Ultimately, the study underscores the need for proactive policy to ensure that AI-driven productivity gains are shared broadly and do not come at the expense of workers’ livelihoods.