Just Three Days of Work... Can AI Fulfill Its Promise?
&cropxunits=450&cropyunits=363&w=600)
Technology leaders' expectations regarding the impact of AI on working hours vary, ranging from reducing them to the continuation of human labor.
Translated by Qiyas from الأنباء — Original source
Amazon founder Jeff Bezos sees the potential to complete five days’ worth of work in three, and to shift family reliance on dual incomes into an optional choice. Anthropic reports that its “Claude” model operates independently for seven consecutive hours, while Mark Zuckerberg argues that smaller teams accomplish more.
JPMorgan Chase CEO Jamie Dimon envisions a workweek of three to three and a half days, with income reaching $300,000 over three decades. Zoom CEO Eric Yuan favors three or four days, Bill Gates imagines two or three, and Elon Musk anticipates voluntary work. Sam Altman, CEO of OpenAI, disagrees, arguing that competition will keep humans working.
Anthropic reduced its relative hiring of junior technical staff, and by August, its AI agents had surpassed employee working hours by a ratio of 6 to 1. A University of California, Berkeley study that tracked employees at an American technology company for eight months documented accelerated work, expanded job responsibilities, and extended hours to verify outcomes. Meanwhile, ActivTrak, analyzing activity data from more than 10,500 users before and after adopting artificial intelligence over 180 days, recorded increased time spent on email, correspondence, and business management.
American annual working hours have declined by approximately 40% since 1870. In 2023, they stood at 2,252 hours in Colombia, 2,228 in Mexico, 2,171 in Costa Rica, 1,953 in Chile, 1,894 in Greece, 1,805 in the United States, 1,528 in Switzerland, and 1,339 in Germany.
Reduced working hours may be accompanied by lower wages and overtime, while it is difficult to measure decision quality compared to attendance hours. The value of trust, relationships, reputation, judgment, and care emerges alongside human judgment, adaptability, and machine oversight.