Historical Comparison Memo
This memo does not build a new forecast. It benchmarks the extreme plausible AI labor scenario against prior U.S. labor-market peaks and shows why the policy burden looks different when reabsorption stays weak.
What it compares
- 1982 recession peak
- Great Recession peak
- Pandemic shock
- Extreme plausible AI case in 2035 and 2040
Main point
On headline unemployment alone, the AI case looks like a severe recession. On support population, participation, and persistence, it looks structurally worse because the burden remains elevated instead of being absorbed within a few years.
Key comparison
- 1982 peak unemployment: 10.8 percent
- Great Recession peak unemployment: 10.0 percent
- Pandemic peak unemployment: 14.8 percent
- Extreme AI case peak unemployment: 10.6 percent in 2037
- Extreme AI case support population: 40.1 million by 2040
- Extreme AI case labor-force participation: 51.4 percent by 2040
Why history understates the risk
Historical peaks were cyclical. This case is framed as structural because task destruction can continue, previously reemployed workers can be displaced again, and wage pressure spreads into remaining human work.