Consensus on AI Capabilities, Growth Uncertain

A large, multi‑group survey shows economists, AI experts, superforecasters, and the public largely agree that AI capabilities will advance dramatically by 2030, yet they sharply disagree on the size and timing of GDP gains. The divergence stems from economists’ emphasis on adoption frictions, capital reallocation to compute, supply constraints (chips, energy, data centers), demographic and geopolitical offsets, and tail risks like social unrest. — If true, policy and investment should focus less on debating whether AI will be powerful and more on managing adoption bottlenecks, supply chains, and social frictions that determine when and whether capability translates into broad economic gains.