A job without a name
Try describing a social-media manager to someone in 1990. The internet made kinds of work ordinary that once had no name. That history makes today’s hopeful promise tempting: AI will create jobs we cannot imagine yet. I hope it does. But a historical analogy is not a job offer.
Change creates winners and losses
Online retail created new work, but it also took business from some physical stores. A study of e-commerce fulfillment centers found declines in nearby traditional retail sales and employment, alongside job gains in other local sectors. Both things can be true: new work appeared, and some workers were hurt.
Exposure is not a layoff forecast
The International Labour Organization estimates that one in four workers worldwide is in an occupation with some exposure to generative AI; 3.3% are in the highest exposure category. Exposure means work could change, not that those workers are about to be fired. Stanford researchers found no broad economy-wide displacement linked to AI, but estimated that employment among 22-to-25-year-olds in highly exposed U.S. occupations was about 19% below its projected path by June 2026. They describe a pattern, not proof AI caused it; weaker hiring appears more important than a surge in layoffs.
Who gets through the new door?
AI may create services and roles we cannot yet name, but we cannot promise how many—or who will get them. Learn the tools and experiment, while paying attention to people whose first career steps may be changing. The internet opened doors and closed others. The question is whether we build enough new doors, and whether the people losing old ones can reach them.