The reasons for regulating AI are far weaker than one could expect.

Comment “Regulation” and we’ll send you an article that explores this further. Warnings about mass job displacement, collapsing entry-level employment, or an economic transformation larger than the Industrial Revolution are still mostly forecasts. So far, the effects of AI are much harder to see in productivity and labor-market data than in headlines. That matters because regulation also has costs. Rules designed around hypothetical harms can protect incumbents, raise barriers to entry, and slow experimentation before we even understand the technology’s real effects. Markets adapt through trial, competition, and learning. Policymakers should respond to demonstrated harms, not regulate first and search for evidence later.