LATEST
AI desk job displacement estimated at 0.5-1% (Feb 2026) Claude writes ~90% of its own code — recursive loop accelerating 50% desk job displacement modeled for 2029-2032 under base case Goldman Sachs: 60-70% of jobs have automatable task components
TOOLS

Interactive: Model AI Displacement
Under Any Scenario

FEB 27, 2026 INTERACTIVE AI LABS TOOLS
MODEL PARAMETERS
How strongly AI self-improvement accelerates the capability curve. 1.0x = no recursion, 5.0x = full recursive loop.
2.5x
Regulatory, cultural, and organizational resistance to AI deployment. 0.1 = frictionless, 1.0 = maximum resistance.
0.50
Months for AI capability to double under standard (non-recursive) exponential growth. Lower = faster base progress.
12 mo
DISPLACEMENT CURVE — REAL-TIME PROJECTION
50% DISPLACEMENT DATE
--
When half of US desk jobs are displaced
DISPLACEMENT AT 2026
--
Current estimated displacement level
PEAK ANNUAL JOB LOSS
--
Highest single-year displacement rate
TIME TO 90% DISPLACEMENT
--
Years from now to near-total displacement
METHODOLOGY

This model uses a logistic (sigmoid) growth function: displacement = 100 / (1 + exp(-k * (year - midpoint))). The growth rate k is derived from the base doubling period and recursive multiplier, then dampened by adoption friction. Specifically: k = (ln(2) / (doublingPeriod / 12)) * recursiveMultiplier * (1 - friction * 0.6). The midpoint shifts based on the effective growth rate — faster growth pulls 50% displacement earlier. This is a simplified analytical model intended to illustrate sensitivity to key assumptions, not a precise forecast. All underlying parameters and formulas are displayed. For full methodology, see our sources page.