Task Automation vs. Job Displacement

Anthropic economists Maxim Massenkoff and Peter McCrory analyzed millions of real Claude conversations to test if AI takes jobs. Key finding: AI handles tasks within jobs but doesn't eliminate roles yet. Studies predict 40%, 60%, or 94% of jobs 'could be automated,' yet this capability doesn't translate to immediate replacement—self-driving cars exemplify how tech lags adoption despite potential.

The 'Could Be' vs. 'Is Being' Gap

Headlines amplify peer-reviewed warnings, but overlook the divide: theoretical automation ≠ real-world displacement. Tech requires integration beyond snaps into existence, preserving jobs like taxi driving despite autonomous vehicle feasibility. This reassures current workers while alarming for career trajectories, as AI shifts demand new paths.

This partial analysis from Anthropic's March 2026 paper urges rethinking AI's career impact beyond panic, focusing on evidence over hype.