AI productivity gains will not automatically shorten the workweek, according to a Financial Times analysis by Adam Shaw. Higher consumption and rising costs in human-intensive sectors could offset potential leisure time.
OpenAI CEO Sam Altman has suggested a four-day work week; JPMorgan Chase CEO Jamie Dimon, three and a half; and Microsoft co-founder Bill Gates, as low as two. However, the analysis notes that these predictions ignore macroeconomic feedback.
Workers benefiting from AI may choose more leisure or higher consumption, but widespread adoption can increase aggregate demand.
This pushes up prices and erodes real wages, forcing some workers to work longer hours to preserve their living standards.
AI boosts supply in automated sectors like software, graphic design, and apparel, lowering prices.
In contrast, human-intensive sectors—childcare, construction, sanitation—experience minimal immediate productivity gains, raising the cost of essential services.
The “Baumol effect,” named for economist William Baumol, occurs when low-productivity sectors must raise wages to compete with high-productivity AI-augmented roles.
🇺🇸 Jamie Dimon, CEO of JPMorgan Chase, said it out loud:
Your kids will work 3.5 days a week. Live to 100. AI is going to cure cancer, stop car crashes, make new materials, save lives.
"Life will be better."
He's not a tech bro dreaming out loud. This man runs the money.… pic.twitter.com/mQUm16nkZx
— Mario Nawfal (@MarioNawfal) May 16, 2026
Historical U.S. data show automation reduced apparel costs but drove up healthcare and housing prices, mirroring today’s projected disparities.
Read: Meta Employee Surveillance Software Tracks Keystrokes to Train AI
Even if AI raises wages, the combined effects of higher consumption and cost pressures could keep or increase working hours, rather than delivering the shorter weeks tech leaders predicted.
The analysis emphasises that AI’s leisure benefits depend on structural economic changes, sector-specific productivity growth, and potential regulatory interventions, rather than technology alone.