technologyliberal

AI at work: more tasks or more jobs?

New York City, USAWednesday, May 6, 2026

< formatted article >

AI: The Double-Edged Sword of the Modern Workforce

From Farm Tools to AI: How Technology Reshapes Labor

History shows that technological advancement doesn’t destroy jobs—it transforms them. When farm tools improved, fewer hands were needed in the fields, yet food grew cheaper, and new roles emerged in burgeoning towns. Electricity followed the same pattern, as did the internet. Now, Dario Amodei, CEO of Anthropic, argues that AI could follow this trend—creating jobs rather than erasing them.

His reasoning hinges on the Jevons Paradox: when efficiency drives costs down, demand often rises. If AI makes lawyers ten times faster, legal services become affordable, and suddenly, everyone wants one. More demand could mean more lawyers, not fewer.

But Amodei introduces a critical caveat: AI moves faster than past technologies. The ATM didn’t wipe out bank tellers overnight—it took decades. AI could compress that timeline from decades to years, leaving workers stranded before new opportunities emerge.

The Hidden Bottlenecks of AI Automation

Even the most advanced AI can’t eliminate human limitations entirely. Amodei invokes Amdahl’s Law: a system’s speed is capped by its slowest component. If AI automates 90% of a lawyer’s workload, the remaining 10%—the nuanced, human-driven tasks—becomes the bottleneck. The job isn’t gone; it’s reshaped, often in ways that favor senior professionals over newcomers.

Wall Street titan Jamie Dimon, CEO of JPMorgan, echoes this sentiment. He contends that capitalism has always shifted jobs, not destroyed them—pointing to agriculture, electricity, and the internet as proof. Yet, Dimon’s optimism glosses over a harsh reality: the economy and the individual worker don’t always align. If AI expands legal services, senior partners thrive while junior associates lose document-review roles.

Retraining, Collaboration, and the Uncertain Path Forward

Retraining programs offer a potential lifeline, but history suggests they’re no panacea. After NAFTA, many displaced workers struggled to adapt—not due to lack of effort, but because the systems in place were ill-equipped to meet their needs.

Amodei and Dimon advocate for business-government partnerships, but past collaborations haven’t always delivered. The real question isn’t whether AI will create jobs—it’s whether companies will prioritize innovation over cost-cutting and whether workers can pivot fast enough to survive the transition.

The future of work isn’t predetermined. It’s a high-stakes balancing act—one where speed, adaptability, and foresight will decide who thrives and who gets left behind.

Actions