When people express concern that AI will replace their jobs, I often think of Jevons' Paradox. It’s a counterintuitive economic principle from 1865, first observed by economist William Stanley Jevons. He noted that as coal-powered steam engines became more efficient, overall coal consumption didn't fall—it skyrocketed. The reason? Efficiency lowered costs, which in turn increased usage and expanded the overall demand for steam power.
I see a direct parallel with AI today. When AI makes professional tasks like writing, coding, designing, and analyzing more efficient, the immediate fear is widespread job loss. However, history suggests that massive productivity gains often have the opposite effect. We saw this pattern with electricity, computing, and the internet. Instead of simply eliminating old jobs, these technologies expanded the economic pie by making new things possible.
As AI increases our productivity, we can expect to see a similar expansion:
- New use cases will suddenly appear.
- More services will become economically viable.
- Entire categories of work will emerge that we can't imagine today.
This leads me to believe the conversation around AI and employment needs reframing. The question may not be whether AI will replace jobs, but rather what new, latent demand its efficiency will finally unlock. History teaches us that efficiency doesn't always shrink an industry; sometimes, it becomes the very engine that scales it.