Anthropic has published an economic model detailing three distinct scenarios for the US labor market through 2030, effectively walking back the catastrophic forecasts broadcast by its own chief executive. Rather than validating apocalyptic talking points, the research grounds the conversation in macroeconomic fundamentals that business leaders can actually plan around.
Under the baseline and moderate scenarios, AI's disruption closely mirrors the rollout of the early commercial internet: white-collar wage stagnation and modest retraining needs without an outright collapse in aggregate employment. Displaced software engineers and customer-support staff face a messy reallocation into skilled trades or healthcare roles, exerting upward pressure on transitional unemployment without breaking the macro framework.
By contrast, CEO Dario Amodei's May 2025 warning that up to 50% of entry-level office jobs could vanish, sending unemployment to 10–20%, gets mathematically filed where it belongs: as an extreme, low-probability outlier. To trigger Amodei's doomsday scenario, the US economy would need to double its GDP every 4.5 years, push knowledge-worker joblessness to 17.9%, and shrink labor's share of GDP from 60% down to 45%. For enterprise executives, the takeaway is clear: Anthropic's research team just provided the analytical rebuttal to its own C-suite hype.